Download as pdf or txt
Download as pdf or txt
You are on page 1of 98

The Business Case for Urban

Small Cells

Commissioned by Small Cell Forum

Issued to: Small Cell Forum


Issue date: 04 February 2014
Version: 3.1
Real Wireless Ltd
PO Box 2218
Pulborough t +44 207 117 8514
West Sussex f +44 808 280 0142
RH20 4XB e info@realwireless.biz
United Kingdom www.realwireless.biz
Real Wireless Ltd
PO Box 2218
Pulborough t +44 207 117 8514
West Sussex f +44 808 280 0142
RH20 4XB e info@realwireless.biz
United Kingdom www.realwireless.biz
Version Control
Item Description
Source Real Wireless
Client Small Cell Forum
Report title The Business Case for Urban Small Cells
Sub title Commissioned by Small Cell Forum
Issue date 04 February 2014
Document status Final

Version Date Comment


3.0 19/01/2014 Complete report issued to Small Cell Forum
3.1 04/02/2014 Addressed SCF member comments

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
About Real Wireless
Real Wireless is the pre-eminent independent expert advisor in wireless
technology, strategy & regulation worldwide. We bridge the technical and
commercial gap between the wireless industry (operators, vendors and
regulators) and users of wireless (venues, transportation, retail and the public
sector) - indeed any organization which is serious about getting the best from
wireless to the benefit of their business.

We demystify wireless and help our customers get the best from it, by
understanding their business needs and using our deep knowledge of wireless
technology to create an effective wireless strategy, implementation plan and
management process.

We are experts in radio propagation, international spectrum regulation, wireless


infrastructures, and much more besides. We have experience working at senior
levels in vendors, operators, regulators and academia.

We have specific experience in LTE, UMTS, HSPA, Wi-Fi, WiMAX, DAB, DTT, GSM,
TETRA – and many more.

For details contact us at: info@realwireless.biz

Tap into our news and views at: realwireless.biz/blog

Stay in touch via our tweets at twitter.com/real_wireless

Copyright ©2014 Real Wireless Limited. All rights reserved.


Registered in England & Wales No. 6016945
Executive summary
Urban small cells are licensed spectrum small cells, deployed by operators in areas of high
demand density and accessible on an open access basis by all the customers of the operator
in that area. They may be deployed outdoors in city centres or parks, often on street
furniture such as lighting columns or on building walls, where they are known as metrocells
or microcells. They may also be deployed in indoor public locations such as retail malls,
airports and railway stations.

In this independent study conducted by Real Wireless on behalf of the Small Cell Forum, we
have examined the business case for urban small cells, drawing on technical and cost
parameters suggested by Small Cell Forum and its members and wider industry sources.

Market drivers

We identified four key market drivers for urban small cells, all of which support the
worldwide growth in both demand and value for mobile broadband services:

1. Capacity enhancement: Ensuring that hotspots of intense demand density can be


served economically.
2. Coverage depth enhancement: Ensuring that acceptable quality and speed of
service is available to users more consistently and deeper inside buildings.
3. Improved user experience: Enhancing the typical speed delivered to users, helping
to meet rising user expectations and application requirements and realising the full
potential of advanced 3G and 4G technologies.
4. Delivering differentiated services: Enabling a platform for location-aware context-
driven services in specific venues.

While the first of these drivers, capacity enhancement, has received most industry
attention to date, we find that the other drivers could lead to a drive for a greater number
of small cells and ultimately greater upside benefits for mobile users and operators alike.

Capacity enhancement

We have conducted a case study comparing forecasts of demand growth with the expected
growth in the quantity of spectrum available and the potential for improved spectrum
efficiency in macrocell networks. The result is a significant shortfall in the capacity from
these sources alone in dense traffic areas, so that greater spectrum reuse via network
densification – i.e. small cells – will be inevitable for most operators. The exact timing and
quantity of small cells needed will differ significantly amongst operators depending on their
existing and planned spectrum network assets. However, it is important to appreciate that
the different sources of capacity (spectrum, technology and topology / densification) act in
combination to give greater gains than any one source in isolation, so operators should
seek advancement in all three simultaneously.

Network densification only makes sense if it can be done cost-effectively compared with
other sources of capacity. The associated costs depend on the details of the existing
network assets, including both the existing macrocells and the range of backhaul options
which are available. Previous studies have not accounted for these in an explicit manner

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
and have tended to assume that the existing network remains static in terms of its own
capacity. Features of our modelling include:

 Explicit account of existing macrocell density and wireline backhaul


 Account of varying wireless backhaul solutions, architectures, performance and
costs
 A detailed model of small cell costs including initial planning and deployment
costs and ongoing operational and replacement costs, validated against industry
data and interviews with key vendors and operators.
 A backdrop of increasing spectrum and spectrum efficiency during the study
period on the existing macrocell network.
 Determination of the required density of small cells to meet a non-uniform
profile of demand across the city based on both coverage capabilities and
capacity

Our base case analysis focused on a dense urban area of 37.6 km2 with LTE traffic growing
at a compound annual growth rate of 35% over the period 2014-2020. We find that the
deployment of small cells to enhance capacity yields a benefit or $48.6m (present value) for
a total cost of ownership of $29.8 million yielding a positive net present value for operators
of $28m and a Return on Investment of 136%. This is therefore a highly cost-effective
investment with a payback period within 4 years. We consider this result conservative, as
the operator bears the costs of initially establishing the small cells sites within the study
period, but these sites will be a useful asset over the long term for the operator.

Nevertheless the specific outcomes depend on the assumed parameters and will differ
significantly amongst operators and markets. We therefore conducted a series of
sensitivity investigations to determine the robustness of the business case, outcomes of
which include:

 The value of the solution increases rapidly with the growth rate of the demand.
Our base case assumed 35% annual growth but many operators face growth at
much higher levels: with a growth rate of 45% the value of small cells increases
by over 260% relative to the base case.
 In the base case, we assumed that the incremental cost of carrying excess traffic
on the macrocells was initially $0.3/GB and increases with the density of demand
to be carried. The small cell layer remains a positive investment proposition even
if the macrocell costs are constant in the face of increasing demand (which is
unlikely) at a level of $0.6/GB. Other studies have suggested costs in areas that
are not heavily constrained by site availability to lie in the range $0.25-$0.7 /GB.
 Whilst placing the small cell at the centre of the hotspot maximises the demand
served, the costs of extending backhaul connectivity to this location can
sometimes outweigh the benefits. In our case study, value was maximised with
small cells around 5 meters off the ‘benefit-maximising location’. In general
though, the business case was not strongly sensitive to this factor, and was
positive over the range of offsets tried.
 The costs of small cell equipment itself are small in the total cost of ownership of
the network (approximately 20% of the overall present value of the costs) during
the study period and costs are dominated by initial site establishment costs (10-
20%) and ongoing opex (approximately 50%). A 20% price increase in the cost of
the small cell results in only a 13% reduction in the net present value of the small

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
cell network. This suggests that operators should seek to maximise the traffic-
carrying capabilities of the small cells deployed as the technology evolves. For
example our base case assumes that the spectrum efficiency of small cells
increases from 0.9 to 1.5 bits per second per hertz over the study period: if that
could be increased to 1.5 bits per second over the whole period then the value of
the small cell deployment increases by over 28% (to $35.8 million).

Overall we consider that the capacity enhancement based market driver is an


important one with a strong business case given carefully planned deployment and this
will be the lead driver for many operators. However the other market drivers have
received relatively less attention but may be dominant for some specific operators in
the near term and for a wider set of operator in the longer term.

Coverage enhancement

Although urban areas are usually well covered by macrocells, there can still be
challenges in achieving adequate coverage reliability and penetration into buildings.
While coverage can be further enhanced by delivery of additional macrocells, in a
dense urban area suitable sites are difficult to acquire and increasingly expensive to
deploy and maintain. We have examined the role small cells can play in improving the
depth of coverage provided by an operator who does not have any low (< 1 GHz)
frequency spectrum and wishes to achieve similar coverage to their competitors.
Depending on the coverage depth and minimum data rate targeted, the costs which
would be incurred to match their competitor via macrocell densification vary from
$30.0m (10 metres, 35.17 Mbps) to $39.4m (15 metres, 32.24 Mbps)1. Achieving
comparable coverage by deploying a dense small cell layer across the same 37.6 km2
city area studied in the capacity enhancement study increases the costs by between
$53.8m and $55.7m. Wide area deployment of small cells is unlikely to be as cost
effective as macrocell deployment. However, if the small cells are targeted more
explicitly at the cell-edge areas where the macrocells deliver unacceptable coverage,
cost savings can be achieved compared to achieving the same target by macro layer
augmentation. These cost savings range from $2.8m to $7.2m.

However this cost saving is only one benefit arising from this deployment and there are
additional benefits to be captured. These include the quality of experience
enhancement in the form of an increase in the median throughput delivered to all
network users (not only those on the small cells). We estimate this benefit will increase
throughput by between 546% and 579% for the 68% to 37.7% of users in the area
targeted by small cells. Based on other work, we estimate this could yield a benefit
ranging from $6.3m to $11.3m for those users via improved working efficiencies and
between $1.0m to $1.9m for the operator. In addition, this coverage augmentation and
performance improvement can be achieved without requiring additional spectrum.
Based on spectrum valuations per population served in Europe and the US, we
anticipate that reuse of the existing spectrum in a small cell layer will save the operator
just over $5m.

Further benefits such as increased capacity and reduced churn are also anticipated but
have not been quantified. Taken together this suggests that targeted coverage

1 Costs are presented on a present value basis.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
augmentation by the use of small cells provides a cost attractive alternative method for
operators to gain high performance coverage in dense urban environments.

Indoor urban small cells

While outdoor urban small cells have received most attention to date, there is also a
need to provide cost effective service to indoor areas with ultra-dense demand
requirements which are often encountered in large public buildings. We have examined
this need via a case study of a busy city centre railway station. We find that the demand
density in such locations may significantly exceed the capacity density of conventional
solutions and urban small cells can play an important role in serving such locations. We
find that the cost of capacity in such a location is between 0.3-2.3 US cents per GB,
which is dramatically lower than the typical cost of macrocell capacity illustrates that
small cells are a highly cost-effective means of carrying capacity and provide flexibility
to respond progressively over time to changing usage patterns.

Overall findings

We conclude that urban small cells, in both outdoor and indoor incarnations are an
essential tool for operators and can yield a positive business case even when an
operator plans to evolve both macrocell capacity and spectrum holdings. However the
timing and quantity of deployment will vary markedly amongst operators who should
carefully determine the appropriate market drivers for them and take full account of
existing network and spectrum assets and demand expectations when planning their
deployments. Nevertheless we expect that a fully-featured small cell layer will be an
essential component of future networks and all operators are advised to secure
appropriate site and backhaul assets in the near term to ensure the efficiency of future
deployments. They are also advised to consider market drivers for small cells well
beyond capacity enhancement in order to maximise their competitive position and
service offering.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
Contents
1. Introduction........................................................................................... 1
1.1 Scope of this study......................................................................................... 1
1.2 About urban small cells ................................................................................. 1
1.3 Where could urban small cells be deployed? ............................................... 3
2. Market drivers for urban small cells ....................................................... 4
2.1 Growth in both demand and expectation ..................................................... 4
2.2 Market Driver: Capacity Enhancement ....................................................... 10
2.3 Market Driver: Coverage enhancement ...................................................... 13
2.4 Market Driver: Quality of experience enhancement .................................. 15
2.5 Market Driver: Value-added service opportunities..................................... 17
2.6 What challenges exist for urban small cells? .............................................. 18
2.7 How does our analysis address these issues? ............................................. 19
3. Sources of capacity .............................................................................. 20
3.1 Demand growth ........................................................................................... 20
3.2 Spectrum growth ......................................................................................... 21
3.3 Spectrum efficiency growth ........................................................................ 22
3.4 The capacity shortfall .................................................................................. 23
3.5 Capacity via indoor small cells in licensed and unlicensed spectrum ......... 24
3.6 Urban outdoor small cells can fill the gap ................................................... 25
3.7 Findings ........................................................................................................ 25
4. City centre small cell business case....................................................... 26
4.1 Anatomy of a small cell network ................................................................. 26
4.2 Business case issues..................................................................................... 26
4.3 Our modelling approach .............................................................................. 28
4.4 Market driver: benefits from avoiding the cost of macrocell capacity and
coverage 29
4.4.1 Cost of macrocell capacity ........................................................................... 29
4.5 Demand ....................................................................................................... 31
4.5.1 Demand and capacity .................................................................................. 31
4.5.2 Macrocell spectrum and spectrum efficiency ............................................. 35
4.6 Small cell radio performance....................................................................... 36
4.6.1 Calculating the number of small cells required ........................................... 36
4.6.2 Radio benefit degradation due to the displacement of the small cell from
the traffic hotspot ....................................................................................................... 38
4.7 Backhaul options and performance ............................................................ 38
4.7.1 Backhaul options ......................................................................................... 38

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
4.7.2 Modelling backhaul performance ............................................................... 42
4.7.3 Backhaul dimensioning ................................................................................ 44
4.8 Small cell cost model ................................................................................... 47
4.8.1 Small cell costs ............................................................................................. 47
4.8.2 Backhaul costs ............................................................................................. 49
4.8.3 Operational and capital costs of different solution types ........................... 52
4.8.4 Core costs .................................................................................................... 54
4.8.5 Summary of costs and comparison with macrocells ................................... 54
4.9 Approach to financial modelling ................................................................. 55
4.10 Capacity-driven deployment case study ..................................................... 56
4.10.1 Capacity-limited scenario description ......................................................... 56
4.10.2 Solution description..................................................................................... 56
4.10.3 Outcomes – base case results ..................................................................... 58
4.10.4 Outcomes – sensitivity analysis ................................................................... 59
4.10.5 Findings ........................................................................................................ 64
5. City centre small cell business case – coverage-driven deployment ....... 66
5.1 Scenario description .................................................................................... 66
5.2 Approach to calculating radio parameters .................................................. 66
5.3 Approach to financial modelling ................................................................. 67
5.3.1 Deriving additional benefit .......................................................................... 68
5.3.2 The cost of 700/800MHz spectrum ............................................................. 70
5.4 Outcomes..................................................................................................... 70
5.4.1 Augment existing macrocell layer ............................................................... 71
5.4.2 Deployment of small cells across the entire study area ............................. 72
5.4.3 Targeted deployment of small cells ............................................................ 73
5.5 Findings ........................................................................................................ 73
6. Transport hub small cell business case ................................................. 74
6.1 Background .................................................................................................. 74
6.2 Case study description................................................................................. 74
6.3 Solution description..................................................................................... 76
6.4 Outcomes..................................................................................................... 78
7. Our findings ......................................................................................... 79
References ......................................................................................................... 82

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
Tables
Table 1: Comparison of demand and capacity growth expectations, 2013-20 ...................... 24
Table 2: Maximum cell range for small cell sites .................................................................. 36
Table 3 Summary of wireless backhaul solution categories ................................................... 40
Table 4: Small cell costs used in this report ............................................................................ 49
Table 5: Backhaul capex and opex for the elements needed to support different cell types 51
Table 6: Summary of costs per small cell type ....................................................................... 52
Table 7: Core network capex and opex (GBP k) for the elements needed to support small
cells 54
Table 8: Total Cost of Ownership for different cell types ....................................................... 54
Table 9: Base case scenario parameters ................................................................................. 56
Table 10: Key financial outcomes for base case ..................................................................... 58
Table 11: Cost elements for low, medium and high cost macrocells (GBP, 2013 values) ...... 67
Table 12: Proportion of low, medium and high cost macrocells ............................................ 68
Table 13: Coverage performance comparison when equal depth of coverage sought ......... 71
Table 14: Coverage performance comparison when equal data rate sought ........................ 71
Table 15: Number of macrocells and associated costs........................................................... 72
Table 16: Number of small cells with associated costs and benefits ..................................... 72
Table 17: Number of small cells with associated costs and benefits ..................................... 73
Table 18: Parameters and assumptions for railway station case study ................................. 76

Figures
Figure 1: Examples of urban small cells .................................................................................... 2
Figure 2: Environments where urban small cells can play a role ............................................. 3
Figure 3: Evolution of mobile broadband devices .................................................................... 4
Figure 4: Effect of increasing mobile data usage per 3G connection on GDP per capita
growth rate [] 5
Figure 5: Global operator survey for Small Cell Forum: percentage of respondents
identifying factors as their top driver for urban small cell deployment (Source: [] ................. 6
Figure 6: The market drivers for urban small cells ................................................................... 6
Figure 7: UK mobile data demand growth vs. revenue growth ............................................... 7
Figure 8: Future mobile demand growth by region (Source: GSMA []) .................................... 7
Figure 9: Long-term UK mobile demand growth (Source Real Wireless []).............................. 8
Figure 10: Mobile demand is strongly non-uniform by time, place and person (Source [53]) 9
Figure 11: Variability of usage density by location ................................................................. 10
Figure 12: Capacity depends on a combination of spectrum, technology and topology of the
network © Real Wireless ........................................................................................................ 10
Figure 13: The rate of increase of spectrum efficiency is reducing over time (Source:
Qualcomm) 11

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
Figure 14: Cooper's Law: Network densification was the dominant source of capacity
expansion from 1950-2000..................................................................................................... 12
Figure 15: Operators accept the need for small cells to meet capacity demands ................. 13
Figure 16: Outdoor small cells can deliver deeper penetration into buildings due to the
angles of arrival of waves into buildings and more consistent coverage due to diverse
coverage 14
Figure 17: Measurements at 2.6 GHz confirm outdoor small cells can provide good indoor
service in many buildings (from []) ......................................................................................... 15
Figure 18: Increase in typical user experience as more small cells are deployed (based on
Small Cell Forum simulations of co-channel 3G outdoor small cells []) ................................. 16
Figure 19: Enhanced indoor service levels from combining small cells with an existing
macrocell deployment (2.6 GHz, LTE) Source: Real Wireless analysis of measurement trial
data [] 17
Figure 20: Global survey of operators: % of respondents citing issues as the top barrier to
deploying outdoor small cells [4] ........................................................................................... 18
Figure 21: Demand growth scenarios over our study period ................................................. 20
Figure 22: Illustrative spectrum scenario, representing the available harmonised
standardised spectrum for downlink traffic over the study period. Mid scenario, with 700
MHz available from 2026 ........................................................................................................ 22
Figure 23: Comparison of spectrum growth scenarios ........................................................... 22
Figure 24: Site spectral efficiency evolution ........................................................................... 23
Figure 25: Analysis of capacity shortfall for varying demand and capacity scenarios ........... 24
Figure 26: Illustrative deployment of urban small cells in a busy city centre ........................ 26
Figure 27: Impact of small cell position with respect to benefit-maximising location........... 27
Figure 28: Our approach to modelling the business case for outdoor urban small cells ....... 28
Figure 29: Cost of capacity rising with demand in an urban area with mid scenario capacity
and spectrum growth for different timing of the availability of low frequency spectrum
(Source Real Wireless, []) ........................................................................................................ 30
Figure 30: Assumed incremental cost of carrying an additional 1GB on the macrocell
network for a capacity-constrained operator versus the demand density to be served on
outdoor cells 31
Figure 31: Illustration of how demand is served by different types of cells. ....................... 32
Figure 32: Indoor and outdoor traffic demand in the UK. ...................................................... 32
Figure 33: Indoor data consumption as proportion of the total data demand ...................... 33
Figure 34: Busy hour throughput served by indoor and outdoor cells .................................. 34
Figure 35: Demand traffic density to be supported on outdoor cells .................................... 35
Figure 36: Benefit degradation vs. placement for case study parameters ............................ 38
Figure 37: Spectrum allocations for terrestrial services potentially suitable for small cell
backhaul 40
Figure 38: Different urban small cell types and their associated backhaul topology used in
this study 41
Figure 39: Variation of LOS probability with distance ............................................................ 42
Figure 40: Variation of throughput versus distance for a particular wireless backhaul
solution 43

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
Figure 41: Illustration of the deployment of small cells ......................................................... 44
Figure 42: Relative proportion of small cells which use street-up versus macrocell-down
backhaul 45
Figure 43: Calculation of obstructed and unobstructed range for a given backhaul
throughput requirement ........................................................................................................ 45
Figure 44: Representative coverage areas of A-type and B-type cells showing where line of
sight (unobstructed) and non-LoS (obstructed) links can be secured back to macrocells or A-
type cells. 46
Figure 45: capex (2013 values) and opex for one year of operation (2013 values) for each
small cell type 53
Figure 46: Calculated number of small cells needed .............................................................. 57
Figure 47: Number of small cells of each type required to support the baseline case .......... 57
Figure 48: Baseline cumulative costs, savings and net cash flows ......................................... 58
Figure 49: Impact of different traffic demand growth rates on net present value ................ 60
Figure 50: Impact of macrocell cost on the NPV .................................................................... 60
Figure 51: Impact of the number of macrocells on the NPV .................................................. 60
Figure 52: Impact of different traffic offload on the NPV ...................................................... 60
Figure 53: Impact of macrocell spectrum efficiency on the NPV ........................................... 61
Figure 54: Impact of small cell spectrum efficiency on the NPV ............................................ 61
Figure 55: Impact of spectrum available on the macrocell layer on the NPV ........................ 61
Figure 56: Impact of spectrum available on small cells .......................................................... 62
Figure 57: Impact of changes on the Benefit-maximising Location on the NPV .................... 62
Figure 58: Impact of changes on wireline NTE cabinet density on the NPV .......................... 63
Figure 59: Assumed relationship between productivity uplift and in-year mobile data speed
improvement factor................................................................................................................ 69
Figure 60: King's Cross Railway Station .................................................................................. 75
Figure 61: Categorisation of ranges of user density ............................................................... 75
Figure 62: Small cells grow in proportion to demand in capacity-limited environment ........ 77
Figure 63: Expenditure associated with initial build and subsequent expansion of small cells
in railway station .................................................................................................................... 77
Figure 64: Incremental cost of capacity on small cell network .............................................. 78

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1
1. Introduction

1.1 Scope of this study

This study sets out to establish the prime drivers for urban small cells and particularly
whether a positive business case for the deployment of such cells exists across a wide range
of scenarios. The work has been commissioned by Small Cell Forum as part of their Release
3 and has been informed by interviews with operator and vendor members of the Forum,
but the contents remain the independent view of Real Wireless.

The work covers a wide range of aspects of costs and benefits from urban small cells, but
places a particular focus on cost and performance variations arising from the trade-off
between radio performance and backhaul cost since these have been seen as an area of
especially significant uncertainty - and scope for innovation - by the small cell community.

Our report is organised as follows:

 In this chapter we explain how Small Cell Forum defines an urban small cell and
explain the environments which stand to benefit from small cell deployment.
 Chapter 2 explains the market drivers for urban small cells and the challenges
which exist in achieving widespread deployments.
 Chapter 3 explains numerically how rising mobile demand leads to a balance of
different traditional capacity-enhancing techniques, but leaves a shortfall which
can only be met by network densification – i.e. small cells. We explore how the
rate of small cell deployment might change under various demand and operator
scenarios.
 Chapter 4 provides an analysis of the business case for outdoor urban small cells
deployed in city centre environments. Two case studies are included – one
primarily driven by capacity, the other by coverage.
 Chapter 6 considers a predominantly indoor deployment of small cells in a
transport hub – specifically a railway station.
 Our overall findings are summarised in Chapter 7.

1.2 About urban small cells

Small cells is an umbrella term covering a wide range of use cases and product
configurations. Within that term the main use cases are for home, enterprise, urban and
rural settings, with many variations. There are no fixed boundaries between these cases but
the main markets drivers are fairly consistent within each use case. Previous work has
examined the business case for home [1] and enterprise [2] small cells.

In the case of urban small cells the deployment environments are characterised by a high
demand density. The small cells themselves may be deployed outdoors on street furniture
(lighting columns etc.) or building walls, typically at heights of several metres above the
ground but below the level of surrounding buildings and macrocells. In that setting such
cells have traditionally been known as microcells but are often also referred to as metrocells
today. Urban small cells may also be deployed in indoor public locations where they may be
referred to as open access picocells or similar. Examples are illustrated in Figure 1.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 1
Figure 1: Examples of urban small cells

The Small Cell Forum provides the following definition of urban small cells:

Urban small cells: This typically refers to small cells that offer capacity for dense
environments, which may be indoor (e.g., shopping malls, transport hubs or stadia) or
outdoor (e.g., parks or city centres) but are clearly urban. Deployment is capacity driven.
There will usually be interaction with the macrocell layer. Urban models are designed for
high traffic areas, these are engineered into robust enclosures suitable for deployment in
unsupervised areas. Although capable of higher traffic capacity of between 16 and 64
concurrent users, these may not require significantly higher RF power because they target a
relatively short range.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 2
1.3 Where could urban small cells be deployed?

Urban small cells can potentially be deployed in a wide range of environments, usually
characterised by allowing general public access, rather than restricting access to a particular
group of people (as may be the case in an enterprise or home location). Such environments
may include:

 Indoor environments such as shopping centres/malls, airports, and concert halls.


 Outdoor environments such as urban parks and city centres.
 Locations with a mixture of indoor and outdoor such as stadiums and railway
stations.

These locations have in common a very high density of users. They also often have
especially demanding usage profiles: for example stadiums during sporting events can
experience higher uplink demand than downlink, as many users wish to share their
experiences via uploaded videos and photographs. Similarly airports and railway stations
may experience periods of unusually intense demand during times of transport congestion.

Figure 2: Environments where urban small cells can play a role

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 3
2. Market drivers for urban small cells

2.1 Growth in both demand and expectation

In recent years, demand for mobile broadband service has risen dramatically, fuelled by
devices providing increasingly convenient internet access. First the mobile broadband
“dongle” provided laptop computers with mobile internet access with adoption
accelerating rapidly from 2006. Starting in 2007 smartphones took off as a mass market
device and showed that personal devices supporting a wide range of apps could drive
significant mobile demand despite small screens. Each successive generation of
smartphone had greater capabilities, and supported richer multimedia capabilities starting
with low rate data, progressing to music and basic video and then to high definition video.
Today devices in a variety of form factors and screen sizes are available, and the diversity of
devices seems set to continue to increase, with individuals often using multiple devices [7].
So demand is increasing in three ways: devices which each consume more data over time,
more people with broadband-enabled devices and more devices per person.

Relative data consumption

Figure 3: Evolution of mobile broadband devices

Access to the services which mobile devices enable brings huge benefits to society;
businesses can conduct transactions more efficiently and speedily, consumers can access
goods and services whenever and wherever they wish to and governments can deliver
public services, from education to healthcare, more efficiently and pervasively than ever
before. As a result mobile broadband availability has gone from being a luxury to a
necessity. The economic benefits of mobile access to these broadband services tend to
grow as data consumption increases, as illustrated in Figure 4.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 4
Figure 4: Effect of increasing mobile data usage per 3G connection on GDP per capita
growth rate [3]

These benefits in mobile usage also bring benefits for mobile operators, who have a new
source of services which make productive use of the 3G and 4G networks and spectrum in
which they have significant investment. However, they also come with significant
challenges for mobile operators. These challenges include:

 Delivering sufficient capacity, at an affordable cost, to ensure that mobile


networks keep pace with the growth in demand: while pricing and data caps
could dampen demand growth, such a strategy risks stifling the economic and
social benefits which would come from such demand for all.
 Ensuring that coverage is available in the places where mobile users require
access. This is increasingly everywhere, including a consistent service deep within
buildings.
 Delivering a quality of experience which meets rising user expectations. Such
quality is a blend of data speeds which rise with time, consistent coverage in the
relevant locations and sufficient capacity to avoid users experiencing congestion.
 Operators could also see the business case improve if differentiated services can
also be delivered for which their customers are willing to pay more.

These are consistent with the results of a recent survey of mobile operators for Small Cell
Forum – see Figure 5.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 5
30

25

% of respondents
20

15

10

Improved coverage
Overall capacity increase

Support location aware services

Reduced network TCO


Support new applications and

Support personalized customer


Filling holes in capacity

Opportunity to integrate Wi-Fi

spectrum (eg high frequency)


Reduced cost of data delivery

Opportunity to harness new


response and services
revenue streams

Figure 5: Global operator survey for Small Cell Forum: percentage of respondents
identifying factors as their top driver for urban small cell deployment (Source: [4]

In summary, the four key market drivers for urban small cells are shown in Figure 6.

Figure 6: The market drivers for urban small cells

All of the above has to be delivered at a cost which makes the associated investment
rational and attractive for the mobile operators. While demand has risen steeply, the
associated service revenue has increased significantly more slowly (Figure 7), driving a
requirement for reducing cost per delivered bit of data relative to the associated revenues
in order to produce a positive business case for additional investment.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 6
Figure 7: UK mobile data demand growth vs. revenue growth

Demand growth is not expected to slow, even in the long term. Figure 8 illustrates that
while projected growth rates vary significantly by region, they are high everywhere. In a UK
example, Figure 9 shows that while central forecasts suggest a demand growth of 80x for
the UK from 2013-2030, there is significant uncertainty, pointing to a need for operators to
also be agile and flexible in the approach they adopt to expanding capacity.

Figure 8: Future mobile demand growth by region (Source: GSMA [5])

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 7
Figure 9: Long-term UK mobile demand growth (Source Real Wireless [7])

The growth in demand on its own does not tell the whole story concerning the
requirements for mobile networks. If the demand were uniformly spread over the whole
area it would be easier to handle. In fact, demand is very strongly non-uniform in several
ways as shown in Figure 10:

 Some users make massively heavier demands on the network than others;
 Some places, notably city centres and indoors, see a very much higher density of
usage than others;
 Some times of day see much higher demand than others and these peak times
are different between urban, suburban and rural areas.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 8
Figure 10: Mobile demand is strongly non-uniform by time, place and person (Source [53])

This means that the demand density during periods of peak demand (measured, for
example, in Mbps per km2) is a more important metric of mobile demand than the total
traffic volume. Locations such as busy city centres, entertainment venues and other
locations accessible to the general public show particularly high demand density, with
densities of up to 1 million users per square kilometre and a high level of activity on mobile
devices. While the majority of usage is indoors, the rate of growth of demand density is
high even in these outdoor locations. Figure 10 illustrates the wide variability of population
densities for a variety of locations.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 9
Figure 11: Variability of usage density by location

2.2 Market Driver: Capacity Enhancement

If operators are to meet this requirement for high and increasing demand
density they need to find solutions which are both cost-effective and
flexible to allow their efforts to be focused on the locations with the
greatest need.

Three generic sources of capacity are available to mobile operators:

 Spectrum, where the total bandwidth as well as the frequency bands and
arrangement of those bands affect the capacity available;
 Technology, where modulation schemes, antenna processing systems and
interference mitigation technique all affect the available spectrum efficiency per
cell;
 Network topology, where the density, location and arrangement of cells all affect
the density with which spectrum can be reused in a given area.

The total network capacity can be expressed as the product of the contributions from these
three sources:
Capacity = Quantity of spectrum x Cell Spectrum Efficiency x Number of cells
[bits per second] [hertz] [bits per second per hertz per [no units]
cell]
Capacity Spectrum Technology Topology

TECHNOLOGY

SPECTRUM Capacity TOPOLOGY

Figure 12: Capacity depends on a combination of spectrum, technology and topology of


the network © Real Wireless

So a gain in any one of these elements acts to enhance the gains in the other two. Since
demand for services is rarely uniform across the area to be served it is usually more
relevant to examine the density of capacity in a small area:

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 10
Capacity density = Quantity of spectrum x Cell Spectrum Efficiency x Cell density
[bits per second [hertz] [bits per second per hertz per [cells per km2]
per km2] cell]
Capacity Spectrum Technology Topology

Spectrum is limited in total quantity. While governments and regulators are working on
ways to bring more spectrum to mobile usage progress occurs over years or decades and
mobile is not the only application demanding extra spectrum. Some spectrum bands are
physically better suited to mobile applications than others. While mobile devices can
support more spectrum bands than in the past the performance and cost increases as the
quantity of spectrum a device can access increases.

Technology capability – expressed as spectrum efficiency – is also limited. The transition


from basic 3G, to its more advanced forms and from LTE towards LTE-Advanced brings
much scope for increased spectrum efficiency, but advanced systems are already nearing
physical limits specified by Shannon’s Law. The rate of increase in spectrum efficiency is
reducing with time, as indicated in Figure 13.

Figure 13: The rate of increase of spectrum efficiency is reducing over time (Source:
Qualcomm)

So it is clear that, while both spectrum and spectrum efficiency will bring capacity gains,
both are finite and increasingly scarce. In economics, scarcity in a resource tends to
increase its cost [6], causing operators to increasingly seek other solutions. Delivering more
capacity with the available spectrum requires that the spectrum is reused in different cells
more efficiently, leading to evolutions in network topology. Denser cells necessarily mean
smaller cells, so small cells are a critical component in future networks.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 11
Using small cells for capacity enhancement is sometimes called offload: however we view
urban small cells especially as an integral part of an operator’s overall heterogeneous
network. The services delivered and overall service quality are intended to be at least
equivalent to the wide-area macrocell network and are a fully managed element of the
network.

Clearly small cells do not come for free: the electronics and associated algorithms are
sophisticated, management systems need to scale up significantly, backhaul needs to be
provided and appropriate sites need to be sourced. The actual number of small cells to be
deployed will be a complex balance of both technical performance factors and comparative
costs between small cells and other approaches which will differ between operators and
markets (we studied this balance amongst cost-effective capacity-enhancing techniques in
detail in [7]). Nevertheless, there is no fundamental limit on the number of small cells
which can be deployed. Indeed, Cooper’s Law suggests that the dominant increase in
capacity in wireless systems worldwide in the second half of the 20th century was indeed
network densification as shown in Figure 14 [8].

Figure 14: Cooper's Law: Network densification was the dominant source of capacity
expansion from 1950-2000

In chapter 3 we will examine the comparative availability of the three sources of capacity to
set a range on the number of urban small cells likely to be required as demand increases.
Clearly, as the cell sizes reduce in future, the number of users served decreases, increasing
the challenge on the cost-effectiveness of small cells. We will examine these cost-related
issues in detail in Chapters 4 and 5 (for outdoor urban small cells) and 6 (for indoor urban
small cells). But mobile operators seem to have already accepted a need for small cells in
increasing numbers, as the following quotations and survey results suggest.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 12
Figure 15: Operators accept the need for small cells to meet capacity demands

2.3 Market Driver: Coverage enhancement

The environments where urban small cells are to be deployed lie


within the footprint of existing macrocells. Coverage at a
geographical level therefore already exists, so this is not the prime
function of urban small cells.

However, there is considerable need for methods to cost-effectively


enhance the depth rather than the extent of coverage. Users demand
a more consistent level of coverage inside the buildings they inhabit
or visit. They are no longer content with moving to a specific part of a
building – such as near a window – in order to make a call, but instead demand consistent
service throughout the building.

The minimum data speed which users consider adequate to provide acceptable service
rises with time: just as dial-up service is no longer adequate for today’s fixed broadband,
early 3G data speeds are now inadequate for the high definition video and responsive
collaboration applications which are used increasingly heavily by consumers.

Given this need, coupled with the increasing tendency of building materials to exhibit
higher building penetration losses as they become more energy efficient, a natural
approach is to install small cells inside buildings and we see considerable scope for this in
enterprise and residential environments. In public buildings, such as railway stations,
shopping centres/malls and airports, indoor urban small cells can also play a role.

However, it is impractical to put small cells inside every potential building needing
enhanced service: the backhaul capabilities may be inadequate and negotiations with every
building owner would be impractical. Most significantly, there is scope for each outdoor
small cell to serve relatively larger numbers of users, enhancing the business case
significantly.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 13
But why not simply serve these locations from macrocells, which would serve even more
users? First, the considerations in the previous section on spectrum and spectrum efficiency
mean that these may not be able to serve all of the potential demand. Second, indoor users
exhibit high propagation losses which place a significant load on the macrocell capacity.
Third, macrocells in many city centres are already at a maximum density given site planning
and interference constraints. So outdoor small cells provide an alternative approach. While
changes to building materials affect both macrocell and outdoor small cell penetration
losses, there is a geometrical difference between them, as illustrated in Figure 16. This
shows that angles of arrival at building walls are different and multiple small cells may
provide a form of diversity2 in reaching locations which would be shadowed by a given
macrocell. As a result, such small cells can deliver good indoor penetration even at high
frequencies as we have verified via measurements (see Figure 17).

Deeper penetration from steeper angles of Shadowed from macrocell,


incidence served from overlapping small
cells

Figure 16: Outdoor small cells can deliver deeper penetration into buildings due to the
angles of arrival of waves into buildings and more consistent coverage due to diverse
coverage

2This form of site-to-site diversity is traditionally called macrodiversity, but is perhaps more appropriately called
microdiversity or even metrodiversity in the context of small cells.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 14
Figure 17: Measurements at 2.6 GHz confirm outdoor small cells can provide good indoor
service in many buildings (from [9])

While the performance is clearly variable according to the distance, angle and construction
of the buildings relative to the small cell location, there is considerable scope for the
outdoor small cells to both enhance the coverage depth and offload a bigger demand level
than might otherwise be anticipated from the macrocell. We expect that every outdoor
small cell will provide some incremental coverage, even where macrocells provide very
solid general coverage3. They may also provide an efficient means of handling mobility
traffic which exhibits many transitions between indoor and outdoor environments, such as
into and out of shops with a low dwell time.

So we see a clear role for outdoor small cells as an intermediate form of coverage depth
enhancement between macrocells for the wide area and indoor small cells for specific
buildings, providing a general enhancement in coverage depth for buildings and locations
which are underserved by macrocells and cannot practically/economically be served by
indoor cells.

We have considered a coverage-driven scenario in our analysis of city centre urban small
cells in Chapter 4.

2.4 Market Driver: Quality of experience enhancement

Most existing analysis of urban small cells focuses on their capacity


potential and rightly so since this is the leading and most immediately
compelling reason to deploy. Such a driver is in a sense a form of
‘pain relief’ to deal with congestion and does not in itself target
benefits for mobile users. However, even when a small cell is installed
with enhanced capacity as the prime motivation, additional benefits
are realised in the form of enhanced services for users. As one
example, Figure 18 shows simulations from Small Cell Forum of the
typical (median) throughput experienced by users when outdoor
urban small cells are deployed in a city centre [10]. With four small cells deployed per
macrocell, the typical throughput across all users – not just those on the small cells -
increases by more than a factor of three. These benefits come from two factors: first, the
cell geometry (shorter distances and better propagation conditions) for small cell users

3There is also considerable scope for outdoor small cells to provide fresh coverage in an economical fashion in
uncovered rural areas, but that “meadowcell” use case is beyond the scope of this report.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 15
improves signal quality for those users and reduces their contention for scarce radio
resources. Second, macrocells become less congested, freeing up resources to better serve
macrocell users. The benefits continue to increase as the number of small cells increases,
although the increase is not linear and will depend in practice on the placement of the
small cells, the interference management techniques employed etc.

600% 523%
Increase in median
throughput 500%
400% 315%
300% 246%
All Users
200% 138%
58%31% Macro Users
100%
0%
1 4 10
Number of small cells per macrocell

Figure 18: Increase in typical user experience as more small cells are deployed (based on
Small Cell Forum simulations of co-channel 3G outdoor small cells [10])

Our analysis of indoor service levels based on 2.6GHz measurements in Figure 19 supports
the general trend of these benefits in practice.

While such benefits may come as a positive side-effect of relieving congestion, there is
clearly also scope for an operator to deliberately target these benefits even in areas where
congestion is not significant. For example, they may be motivated to deploy over a wider
area, for example into suburban areas rather than only in the urban core.

In short, an operator deploying a widespread network of outdoor urban small cells would
be in a position to offer their customers higher typical speeds than their competitors,
because their network included more small cells than their competitors. This creates a new
and potentially compelling motivation for deploying small cells. Whether there is a business
case for such a deployment is the subject of our coverage-driven case study in chapter 4.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 16
Figure 19: Enhanced indoor service levels from combining small cells with an existing
macrocell deployment (2.6 GHz, LTE) Source: Real Wireless analysis of measurement trial
data [9]

2.5 Market Driver: Value-added service opportunities

Small cells could enable value-added services based on their


ability to deliver high geographical precision and hence rich
location, context and presence data to applications via
standardised application programming interfaces controlled
by operators, enabling an open ecosystem for developers
while enabling operators to monetise access to these
capabilities.

While outdoor small cells deployed for public access may not
lend themselves to a differentiated service proposition, urban
small cells are also suitable for venue-specific installation,
where there is clear opportunity particularly in indoor retail
locations. Mobile analytics could provide footfall data based
on rich location capabilities, helping to monitor and enhance the use of retail space for the
retailers and property managers, while also providing targeted offers and informational
content for mobile users on an opt-in basis. Retail locations are not limited to shopping

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 17
centres/malls: indeed airports, railway stations, sports and entertainment venues all
provide a wide range of retail outlets.

This is only one example of many possibilities for small cells to provide enhancements.
Many others are described in Small Cell Forum publications [11]. While these are
speculative at the current time, it is expected that they will provide additional upside on
the urban small cell business case over time, provided the ‘core’ business case for coverage,
capacity and user experience stacks up. Some indication of revenue opportunities from
such services has been factored into our analysis of a transport hub (railway station) in
Chapter 6.

2.6 What challenges exist for urban small cells?

Given the market drivers for small cells outlined above, urban small cells appear to be an
attractive complement to existing macrocell technologies for operators, creating a flexible
heterogeneous network or HetNet which can be tailored to meet varying demand.

However, while the concept of microcells has existed for at least twenty years, there are
significant challenges in widespread deployment which must be addressed. A survey of
operators’ views, commissioned by Small Cell Forum, on these challenges is summarised in
Figure 20.

30

25
% of respondents

20

15

10

0
Cost of equipment

Interference
Site costs

Site acquisition
Securing optimal site locations

Monetization

Power availability and cost


Backhaul availability and cost

Network provisioning and

Macro network interworking


management

Figure 20: Global survey of operators: % of respondents citing issues as the top barrier to
deploying outdoor small cells [4]

The key issues which emerge many times over are as follows:

1. Deployment complexity: With issues such as zoning/planning permissions, public


health & safety concerns, and power, weight and wind loading all contributing to
this complexity and uncertainty.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 18
2. Backhaul: In principle small cells can make better use of any existing backhaul
technology than traditional cells. However the availability of fibre in many cities is
not yet sufficient to provide the basis for all small cell deployments and the fibre
termination locations may not be in the right locations to optimise the small cell
benefits, necessitating wireless backhaul links to bridge the gap. The specific
choice and mixture of backhaul technologies will thus depend on many factors.
3. Business case: While small cells may be essential to meet the largest level of
capacity density in future, the extent of deployment will depend on a positive
business case. The cost of delivering each additional data byte must reduce. If
the costs are too high, operators will have to find other ways to manage demand
growth, including charging and policy control to dampen demand. If on the other
hand a cost-effective approach is available, operators may be motivated to
deploy small cells over a wider area.
4. Management and interoperability: Most operators seem convinced that outdoor
small cells can work and technical studies plus standardisation activities are in
progress to ensure this. However, the processes of managing large numbers of
small cells working in close harmony with the macrocell network has not yet been
widely tested. Likewise some operators and vendors see a need to ensure open
interoperability so that small cells and macrocells in a given area can be mixed
between operators and vendors without excessive performance sacrifice.

Our focus in this study is on the business case aspects. However other aspects, notably the
availability of suitable backhaul solutions, may play a significant role in determining costs
and performance trade-offs so we have also factored these into our modelling. Other
aspects of these challenges are being addressed by Small Cell Forum within their Release 3
[12].

2.7 How does our analysis address these issues?

We have chosen to examine the motivations for urban small cells and the associated
business case via three sources of analysis:

 Demand vs. capacity analysis: A fundamental demand vs. capacity analysis in


chapter 3 points to a need for substantially increasing density of cells, making
clear that smaller cells will be needed, although the rate of increase depends on a
balance versus other sources of capacity and the associated costs.
 City centre deployment case studies: An investigation of city centre deployments
of outdoor urban small cells in chapter 4 considers the business case explicitly by
comparing costs against benefits and examines how the balance of these changes
according to differing macrocell network density and backhaul availability. Two
case studies are considered, along with various sensitivities:

1. Case study a: Capacity-driven


2. Case study b: Coverage depth-driven

 Transport hub case study: A busy transport hub is examined in chapter 6 to


illustrate deployment of indoor public-access small cells.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 19
3. Sources of capacity
As described earlier, there are three basic sources of additional mobile capacity: more
spectrum, greater spectrum efficiency through technology gains and greater spectrum
reuse through network densification and topology evolution. This chapter provides an
indicative study of the relative numerical impact of these sources to determine to what
extent network topology changes are essential. Our study focuses on the period from 2014-
2020 and makes use of a previous detailed study: although this study was conducted in the
context of the UK many of the general findings are expected to be consistent with other
markets.

We will conclude that, despite spectrum increases, performance increases on the


macrocells and an increasing level of indoor offload, there is a shortfall of capacity which
can only be met via urban small cells.

3.1 Demand growth

The demand forecasts in Figure 9 were created based on a set of detailed scenarios which
examined growth of individual device consumption, allocated by person, place and time.

The mid demand scenario amounts to some 79x over 2012-30, or 10x over our study period
of 2013-2020. This is equivalent to a compound annual growth rate of 39%. The high
growth scenario amounts to some 297x over 2012-30, or 19x over our study period of
2013-2020. This is equivalent to a compound annual growth rate (CAGR) of 42%. Both are
illustrated in Figure 21.

Figure 21: Demand growth scenarios over our study period

Although these both represent substantial growth, it is clear that even higher growth is
expected in some regions and by some observers when compared to the 54, 56% and 76%
CAGRs for Europe, North America and Asia-Pacific quoted in the study for GSMA in Figure 8
over a slightly shorter time frame (2012-16).

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 20
3.2 Spectrum growth

Spectrum is a basic necessity for wireless systems. The capacity of a wireless system
depends directly on the quantity of spectrum available and increasing the amount of
spectrum can be a very cost-effective means of increasing capacity. However such
spectrum does not come for free: as well as the cost of initially acquiring the spectrum and
maintaining ongoing license fees, there are costs in upgrading network equipment and
ensuring mobile devices support the relevant spectrum bands.

The dominant source of spectrum for mobile applications is spectrum provided on a


licensed, exclusive basis. Suitable spectrum needs to be internationally harmonised to allow
attractive mobile devices and it needs to be included in standards, typically those produced
by 3GPP. It also needs to have appropriate physical properties to yield adequate coverage
and capacity for practical antenna sizes and powers. In practice this tends to limit useful
mobile spectrum in the range 450 MHz – 3.5 GHz.

Given these constraints, there is clearly a limited supply of spectrum for mobile
applications. Regulators internationally are engaged in locating additional spectrum for
mobile broadband: for example in the US there is a government commitment to auction
500 MHz of spectrum drawn from federal and commercial sources [13]. The UK
government announced in 2010 a similar commitment to find at least 500 MHz of public
sector spectrum below 5 GHz over the next ten years for mobile communication uses [14].
The European Commission’s Radio Spectrum Policy Programme published in 2012 includes
an action to ensure that at least 1200 MHz spectrum is identified to address increasing
demand for wireless data traffic [15]. In 2015 the World Radio Conference will consider a
specific agenda item to identify additional frequency bands suitable for delivery of mobile
broadband applications.

Despite these initiatives, however, the actual quantity of additional spectrum is limited and
the timescales involved are long. Figure 22 provides an illustration based on a UK study of
the scope for additional harmonised spectrum, representing an increase of some 11 times
between 2012 and 2030. Although this growth is substantial, a large proportion is
accounted for by spectrum which was initially assigned for 4G use in early 2012, beyond
which the growth rate is much reduced. Over the study period of interest the increase in
spectrum availability is only a much more modest 82% from the end of 2013 to the end of
2020 in the mid scenario and even in the high scenario the growth is only 86%. These are
shown in Figure 22 and are equivalent to annual growth rates of 8.9% (mid), 9.3% (high)
and 6.9% (low), far short of the projected 39-42% demand growth rates. These also relate
to the mobile market as a whole: an individual operator may need to make the best of the
spectrum they have in hand at the start of this period for the entire period, having no
certainty of any additional spectrum.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 21
700
3600 MHz -3800 MHz

Available net downlink spectrum (MHz)


3600 MHz -3800 MHz
600
2600MHz (unpaired)

2300 MHz (2310 - 2390 MHz)


500
2100 MHz TDD (2010 -2025 MHz)

2100 MHZ TDD (1900 - 1920 MHz)


400
1452 -1492 MHz

3400 MHz - 3600 MHz


300
3400 MHz - 3600 MHz

2600 MHz
200
2100 MHz

1800 MHz
100 900 MHz

800 MHz
0 700 MHz in 2026
2010 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030

Figure 22: Illustrative spectrum scenario, representing the available harmonised


standardised spectrum for downlink traffic over the study period. Mid scenario, with 700
MHz available from 2026

Figure 23: Comparison of spectrum growth scenarios

3.3 Spectrum efficiency growth

There are many ways in which existing macrocells can be upgraded to increase the capacity
they can deliver in a given quantity of spectrum – i.e. their spectrum efficiency.

Potential sources of improvement in spectrum efficiency include:

 Increased modulation and coding efficiency (e.g. to 64 QAM from QPSK)

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 22
 Increased sectorisation (e.g. from three- to six-sector)
 Increased MIMO order (e.g. from 2 to 4 elements or more)
 Introduction of Coordinated Multipoint Transmission
 Introduction of Carrier Aggregation

All of these techniques can potentially increase capacity. We have considered the rate of
spectrum efficiency growth over time which might arise from these techniques, taking into
account factors such as the rate of availability of supporting features on mobile devices,
constraints on site upgrades etc. Figure 24 illustrates these resulting spectrum efficiencies
for several growth scenarios.

Figure 24: Site spectral efficiency evolution

3.4 The capacity shortfall

The anticipated growth in demand and capacity is compared in Table 1 with capacity
growth in the case of no network densification. The expected growth in demand of
between 10.2× and 19.0× is not matched by the growth in capacity in either the mid (6.65×)
or high (7.78×) cases. Figure 25 shows how the shortfall in the available capacity growth
relative to the demand growth under various scenarios of the spectrum and spectrum
efficiency growth. This points to a clear need for network densification.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 23
Table 1: Comparison of demand and capacity growth expectations, 2013-20

Parameter Low Medium High


Demand growth 2013-2030 (factor) 10.2 19.0
Demand growth 2013-2030 (CAGR) 39% 42%
Spectrum growth 2013-2030 (factor) 2.2 2.6 2.8
Spectrum growth 2013-2030 (CAGR) 6.9% 8.9% 9.3%
Spectrum efficiency growth 2013-2030 (factor) 2.21 2.58 2.79
Spectrum efficiency growth 2013-2030 (CAGR) 12.0% 14.5% 15.8%
Capacity growth absent densification (factor) 4.89 6.65 7.78
Capacity growth absent densification (CAGR) 19.7% 24.7% 26.5%

Figure 25: Analysis of capacity shortfall for varying demand and capacity scenarios

3.5 Capacity via indoor small cells in licensed and unlicensed spectrum

One means of network densification is to carry some traffic on dedicated indoor cells. Given
that the majority of mobile traffic is created and consumed indoors, this is an efficient
means of reusing precious spectrum without placing excessive load on outdoor cells. Indoor
solutions may include residential and enterprise femtocells and picocells, traditional indoor
solutions such as DAS and Wi-Fi systems operating in unlicensed spectrum. Such an
approach is often referred to as ‘offload’, although we prefer to avoid the use of this term
given its potentially negative connotations for the value of the traffic and the associated
user experience.

Although these indoor cells are an attractive and growing approach to serving indoor traffic
and can generate an attractive business case for particular locations, [1], [2], it is clearly
impractical to serve all the traffic in every building in this way, as it is unlikely that every
building will have small cells for every operator due to cost, backhaul and access (rental
cost, agreements with site provider) limitations. Likewise, not all areas of high traffic

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 24
density are indoors, including busy shopping streets, sports stadia, entertainment festivals
and transport hubs. These are exactly the areas targeted by urban small cells.

3.6 Urban outdoor small cells can fill the gap

Thus we expect urban small cells to fill an important gap in meeting traffic demand in
locations where the demand density is too high to meet via macrocell solutions and in
places where indoor solutions alone are insufficient to meet the needs.

However, whilst this indicative, high-level analysis provides a view of the relative growth of
demand and supply it is overly simplistic in several ways:
 It takes no account of the relative costs of adding capacity via technology,
spectrum or topology options. Even in cases where there is a sufficient capacity,
the presence or absence of given spectrum bands could significantly affect the
expenditure required to upgrade the capacity of existing sites at the required
rate.
 It accounts for capacity in only an aggregate form and does not reflect the
potential for very local hotspots of traffic to overload or degrade the quality of
the network.
 It does not reflect the likelihood that future services will require increasing typical
data rates as well as quantities which could require network expansion at a rate
which would be strongly influenced by the available spectrum.
 It treats all spectrum as equal in its impact and does not reflect the varying
coverage and interference properties amongst bands.

We will address these issues in our detailed analysis in later chapters.

3.7 Findings

It is clear that while spectrum and spectrum efficiency growth on existing macrocell sites is
essential, it is not sufficient to meet the demand growth in the long term and could indeed
lead to capacity crunches in the nearer term where the combination of improved spectrum
efficiency and available spectrum may not be sufficient to meet demand growth. In these
circumstances, and in particular areas with access to the general public, network
densification via urban small cells is expected to be an essential tool for operators. The
number and timing of the requirement may vary substantially for a given level of demand
growth according to an operator spectrum and network assets.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 25
4. City centre small cell business case

4.1 Anatomy of a small cell network

Figure 26: Illustrative deployment of urban small cells in a busy city centre

Urban small cells (i.e. metrocells or microcells) will be deployed in busy outdoor locations,
particularly city centres. They are likely to be deployed on street furniture sites such as
lighting columns (lamp posts) and CCTV poles, or else placed on the side of building walls,
typically at heights from 5-12m above ground. Unlike macrocells, they are deployed below
the rooftop heights of surrounding buildings, enabling them to focus capacity where
needed and control inter-cell interference.

Such cells will be backhauled via a variety of means: a form of wireless link may connect
them back to macrocell sites, where they will share the existing macrocell backhaul. They
may also be connected directly or via a wireless solution to street-level wireline via copper
or fibre. The integration of small cells to the macrocell network may be directly into the
core network or via a gateway.

4.2 Business case issues

Previous studies, including by Small Cell Forum [10], have shown that when small cells are
optimally located they can generate significant capacity benefit at the same time as
improving the typical user experience in terms of the consistent and speed of service.
However, there are several challenges in the business case which need to be addressed,
including the following points:

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 26
 It will not always be possible to locate a small cell at the point which maximises
its radio network-enhancing benefit, since the quantity and location of the
demand may be uncertain and since existing street furniture may not be available
at the benefit-maximising location (BML).
 The topology of urban small cells means that backhaul needs to be extended
much closer to the user and backhaul solutions need to be appropriately cost-
effective for this. While all backhaul traffic capacity is ultimately provided by a
wireline link into the core network, the existing density and cost of wireline
solutions means that most small cell installations will also need some element of
wireless interconnect close to the RAN network elements. The costs associated
with this extra backhaul layer need to be fully considered in order to establish the
existence of a positive business case. This in turn depends on the details of the
capabilities of the backhaul solutions considered.
 In order to reduce costs associated with backhaul, it may be advantageous to
deliberately move small cells from the benefit-maximising locations towards
locations of wired backhaul or a wireless line of sight availability to reduce costs.
This will however increase the required number of small cells to achieve a given
benefit, again calling into question the overall cost-effectiveness. This effect is
illustrated in Figure 27.

Radio benefit

Demand hotspot
Smallcell
coverage

Radio benefit

Backhaul capacity

Backhaul & site cost

Figure 27: Impact of small cell position with respect to benefit-maximising location

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 27
4.3 Our modelling approach

Our modelling approach is illustrated in Figure 28 and described at high level below. Details
are provided in the following sections.

Figure 28: Our approach to modelling the business case for outdoor urban small cells

In the Market Drivers section of our model, we specify the benefits which the operator
seeks to gain from the use of small cells. For our first case study we examine an operator
who faces macrocell capacity constraints and seeks a more cost-effective means of serving
additional demand. In another case study we examine an operator with spectrum
constraints on macrocell coverage who seeks to reduce the cost of achieving adequate
coverage but also hopes to gain additional benefits from enhanced user experience.

In the demand section we specify the traffic demand density over the study area of
interest, typically a dense area of a city. We also specify the existing macrocell network
density, the available spectrum and the opportunity for macrocell upgrades to enhance
their capacity. The incremental demand remaining after this macrocell capacity is
exhausted is the demand which must be served by the small cell layer.

The radio section of our model specifies the radio access network performance available
from small cells, based on their maximum coverage area and the available capacity per cell.
It also specifies how the performance degrades as the small cell location is moved away
from the BML. Taken together these factors can then be used to determine the number of
small cells required to meet the incremental demand.

The backhaul module specifies the density of existing wireline backhaul, separated into
existing macrocell backhaul locations which are suitable for carrying small cell demand and
the street level locations such as fibre termination points. The module contains the details
of the performance of several wireless backhaul solutions to determine the viable forms of

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 28
wireless connectivity to backhaul small cells which are not directly connected to a wireline
solution.

The cost module assigns both capital and operational costs to every element of the small
cell network. These unit costs are then multiplied up according to the number of small cells
and the form and nature of backhaul to yield the total costs incurred in any given time
period.

All of the modules described allow the demand and solution parameters to vary with time
over a given study period. In the case of our study, the period of interest is assumed to be
2014-2020 and all parameters are calculated for every year over that period.

Finally financial analysis is conducted to compare the benefits and the costs over time and
compute several key financial metrics, including total cost of ownership, net present value
and return on investment. These are examined to determine the attractiveness of the
business case.

4.4 Market driver: benefits from avoiding the cost of macrocell


capacity and coverage

When the operator’s market driver for small cells is capacity enhancement, the benefits of
the small cell layer derive primarily from a reduced cost to serve traffic with a given quality
compared to the equivalent service on the macrocell network. In some cases it may be
essentially impossible (irrationally high cost or physically implausible) to serve traffic via the
macrocell, e.g. where the macro capacity is exhausted or where the required coverage
depth cannot be achieved using only high frequency spectrum. Nevertheless for most
operators their initial option for capacity enhancement will be to upgrade existing
macrocell sites, since these sites represent a substantial investment with associated sunk
costs.

Note that considering only the capacity delivered by small cells is a conservative estimate of
their benefit – there will be further upsides for user performance and coverage depth that
are not considered in this capacity-only analysis.

4.4.1 Cost of macrocell capacity

The cost of providing incremental mobile traffic (cost per GByte of data) from macrocells
varies according to the environment, the existing macrocell capabilities and the quantity
and type of spectrum available to an operator. Typically rural environments can require
more costs associated with establishing civil infrastructure and provision of utilities, and
urban environments can have higher rental costs in some prime locations. Operator co-
location and use of more spectrum and multiple technologies at the same site can reduce
the cost per GB compared to a new site built to augment coverage. In some dense urban
areas, e.g. London, macrocell site density can be as high as 10 sites/km2 [7] and adding
further macrocells for capacity can be problematic on both interference and site acquisition
grounds. This raises the cost in these capacity limited environments.

Based on the various factors above, we should not be surprised then that there is a wide
range of values on the cost of providing 1GByte of data to subscribers using macrocell
networks from different industry sources. Ericsson [16] estimate the cost of providing
1GByte using HSPA+ on a well utilised network to be less than 1 Euro and quoted an

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 29
operator as confirming that costs of delivering 1GB is below 2 Euro (1.8 GBP). In low traffic
areas, Plum assessed the costs of providing 1Gbyte [17] on an LTE network to drop from
just above 3 Euro in 2011 to 0.5 Euro in 2014, flattening off to reach a cost of approximately
0.1 Euro by 2020. Plum estimated that lower costs would occur sooner in high traffic areas.
In 2010, a study for Small Cell Forum [18] estimated the future costs of delivery of 1Gbyte
on a capacity constrained macro network to be US$4.80 (GBP 2.97, or Euro 3.56) and
approximately US$3.30 (GBP 2.05, or Euro 2.45) on a coverage constrained network.

However, none of these previous studies gives a clear relationship between the network
parameters, the demand density and the associated costs. Our own previous study of
capacity factors looked at this relationship directly but did so for the case where spectrum,
spectrum efficiency and site density were all able to increase significantly over time and
related to the entire mobile market rather than to a particular operator (Figure 29).
Nevertheless this shows how costs can accelerate as demand density rises, sometimes very
rapidly when a capacity ‘crunch’ occurs when demand growth exceeds the capability of
existing sites.

£6.0
Urb Lon, 2026
Millions

£5.0 Urb Lon, Never


Urb Lon, 2020
£4.0 High demand Mid demand

£3.0

£2.0

£1.0 Low demand


£0.0
2025
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024

2026
2027
2028
2029
2030
Figure 29: Cost of capacity rising with demand in an urban area with mid scenario
capacity and spectrum growth for different timing of the availability of low frequency
spectrum (Source Real Wireless, [7])

Although there are many variables to consider, we believe it is reasonable and consistent
with the above studies and the discussion in chapter 3 to assume that demand density is
the major driver for the increasing cost of capacity. However the relationship will not be
linear due to the progress made in macrocell technology, price erosion of wide area
solutions and the potential for the operator to acquire additional spectrum.

For the base case in our study we have assumed that cost of supporting 1 GB on the
macrocell layer will grow over the study period from a cost of 0.2 GBP in 2014 . From that
base it rises as function of the demand density (net of demand served by indoor cells)
resulting in the cost per GB of supporting traffic in the capacity constrained macro layer as
shown in Figure 30. However, given the wide range of factors which affect these values we
have also conducted sensitivity analysis to determine the level to which macrocell capacity
costs would need to fall to be more cost-effective than small cells.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 30
Cost to carry 1GB on macrocell site
1.2

0.8
Cost [£/GB]
0.6

0.4

0.2

0
2014 2016 2018 2020 2022 2024 2026 2028 2030
Year

Figure 30: Assumed incremental cost of carrying an additional 1GB on the macrocell
network for a capacity-constrained operator versus the demand density to be served on
outdoor cells

4.5 Demand

4.5.1 Demand and capacity

Figure 31 shows how user demand is partitioned between indoor and outdoor sources with
the majority of demand typically being generated indoors. The figure also shows how we
assign that demand to different sources of capacity with indoor cells serving a portion of
the indoor demand and outdoor cells serving both outdoor and remaining indoor demand4.

4We note that some outdoor demand could also be served by indoor cells in an ‘inside out’ model, but we have
not accounted for that possibility here.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 31
Figure 31: Illustration of how demand is served by different types of cells.

The demand for data traffic is generated from a combination of indoor users and outdoor
users. For instance, a forecast of the total data traffic demand in the UK is shown in Figure
32 based on a previous Real Wireless study for Ofcom [7]. This aligns closely with the ‘mid’
demand scenario analysed in chapter 3.

Total demand
Indoor (domestic and business) demand Outdoor demand

600,000

500,000
Demand [TB/month]

400,000

300,000

200,000

100,000

-
2014 2015 2016 2017 2018 2019 2020
Year

Figure 32: Indoor and outdoor traffic demand in the UK.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 32
According to [19], around 75% of data use is indoors today. Indoor usage is expected to
increase up to 90% by 2020 [7].

Figure 33 shows our assumed increase in indoor data consumption.

Indoor (Home+Office) demand 95%

90%

85%

80%

75%

70%

65%

60%
2014 2015 2016 2017 2018 2019 2020
Year

Figure 33: Indoor data consumption as proportion of the total data demand

A proportion of this indoor traffic is “offloaded” onto fixed networks using indoor small
cells, including Wi-Fi, femtocells and picocells in residential and enterprise settings. We
assume offload grows at a linear rate over the study period based on a review of various
studies in [7].

The remaining indoor traffic which is not offloaded to indoor small cells, together with the
portion of traffic coming from outdoor i.e. outdoor demand, all has to be served by outdoor
cells. The outdoor cells can be either macrocells or outdoor small cells. We calculate the
demand served by the macrocells based on the density of existing macrocells in the study
area and the capacity served by a macrocell. It is usually difficult to find and acquire
suitable site locations to deploy macrocells in densely populated areas such as the area
selected in this study. Therefore, the opportunity to increase the macrocell site count is
very limited. In this study, the growth in macrocells is assumed to be 1% during the study
period in line with [19]. We assume that the operator of interest in the study has a market
share of 40%.

The busy hour throughput for one operator is calculated based on the following
assumptions [20]:
 Busy hour utilization of available RAN capacity is 80%.
 Busy hour share of the daily traffic is 20%

In addition, the following parameters have been used during this study:
 Area: 37.6 km2

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 33
 Number of macrocell sites: 130 [21]

These values were taken from the four most densely boroughs of Central London
incorporating the City of London, City of Westminster and Kensington and Chelsea but
could apply to any city with a similar density of demand and macrocells. The study area is
partitioned between dense and less dense areas based on the traffic demand. Areas with a
higher traffic demand are known as “hot spots”. Macrocell capacity is computed in a similar
way to chapter 3, considering changes in the spectrum efficiency and spectrum availability
over the study period.

Once the demand served by the macrocells is found, the demand which needs to be served
by outdoor small cells is the remainder of the outdoor demand.

Figure 34 shows the busy hour throughput served by indoor and outdoor cells for our base
case study.

Indoor and outdoor traffic


Capacity served by outdoor macrocells
Capacity served by outdoor smallcells
Capacity offloaded to indoor cells[Gbps]
80 5,000
by outdoor cells [Gbps]
BH throughput served

by indoor cells [Mbps]


BH throughput served
60
4,086 4,000

2,855 3,000
40
1,992 2,000
20 1,387
963 1,000
462 668
- 0
2014 2015 2016 2017 2018 2019 2020
Year

Figure 34: Busy hour throughput served by indoor and outdoor cells

Another way to consider the traffic is to look at the demand traffic density that is to be
supported on outdoor cells (both macro and small cells). As this demand density increases
the difficulty of continuing to support this traffic exclusively on the macrocell layer
increases. This is shown in Figure 35 and is used as the driver for the costs which would be
incurred if small cells are not used via the relationship of Figure 29.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 34
Demand density on outdoor cells
140,000

Demand Density (Mbps/Km^2)


120,000

100,000

80,000

60,000

40,000

20,000

0
2014 2015 2016 2017 2018 2019 2020
Year

Figure 35: Demand traffic density to be supported on outdoor cells

4.5.2 Macrocell spectrum and spectrum efficiency

As explained in the above section, an operator with a limited amount of spectrum is


selected for the study. The operator is assumed to have acquired or refarmed 40 MHz of
spectrum at the start of the study period and is not able to further increase it during this
period. Figure 35 shows the capacity and spectrum efficiency of macrocell (3 sector) and
small cell (single sector) sites based on analysis in [19].

Macrocell site capacity [Mbps] Smallcell site capacity [Mbps]


Macrocell site SE [bps/Hz] Smallcell sites SE [bps/Hz]
4.00 160.00

3.50 140.00
Spectrum efficiency [bps/Hz]

3.00 120.00 Capacity [Mbps]


2.50 100.00

2.00 80.00

1.50 60.00

1.00 40.00

0.50 20.00

0.00 0.00
2014 2015 2016 2017 2018 2019 2020
Year

Figure 35: Spectrum efficiency and capacity of macrocell and small cell sites

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 35
The capacity supported by macrocell sites and small cell sites can be calculated by
multiplying spectrum efficiency by the amount of spectrum used in the site.

4.6 Small cell radio performance

This section explains the radio performance characteristics assumed for small cells.

4.6.1 Calculating the number of small cells required

The number of small cells is calculated by assuming that they are required to serve the total
demand in the hotspot areas which is not served by macrocells. We check both that the
coverage area of the small cells is enough collectively to serve each hotspot area and also
that their total capacity is sufficient to serve the hotspot demand. We also check that their
total capacity, incorporating a degradation for being placed in a location which is not the
benefit-maximising location, is sufficient to serve the demand in the hotspot. The actual
number needed is the greater of these coverage-driven and capacity-driven results.

The maximum small cell coverage (in the absence of capacity constraints and for dedicated
spectrum deployment) can be calculated using the cell ranges given in Table 2 based on
[22].

Table 2: Maximum cell range for small cell sites

Type Cell range (m)

Small cell 50

Small cell sites are assumed to be single sector sites. This is the reason for the large
difference in spectrum efficiency of a macrocell site and a small cell site in Figure 35.
Further, as the technology evolves, the spectrum efficiency of both macrocell sites and
small cell sites increases over time. The spectrum efficiency of a small cell is assumed to be
24% larger than a single macrocell sector at the relevant point in time to account for their
improved geometry factor, based on our review of simulation results (See A2.9 of [7],
source Real Wireless).

The number of small cells required per macrocell can vary depending upon various factors
such as inter site distance between macrocells and small cells, antenna radiation pattern at
small cells and the distribution of user devices within the coverage area of macro cells.

In the shared carrier deployment adding small cells also add interference to macrocell
users. However, the degradation in throughput of the macro cell users is very small
compared to the overall capacity improvement [23]. With the deployment of LTE-Advanced
small cells at least 180% increase in the capacity is expected to be achieved [24].

From the perspective of operators deploying small cells the amount of available spectrum
has crucial impact on the number of macro and small cells. For example a Nokia-Siemens
study [25] shows that for a capacity constrained operator the number of small cells
required to meet the increasing traffic demand grows very rapidly. However, if the operator

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 36
had an additional carrier this will balance the number of macrocells and small cells and
reduce their counts significantly.

An Ericsson study of the user throughput performance for macrocell and small cell users
with the deployment of additional LTE small cells is reported in [26]. It assumes a hotspot
scenario where 2/3 of total users are within the hotspot area and that no enhanced
interference management schemes are employed i.e. 3GPP Rel. 8/9 only. The results show
that increasing the number of small cells substantially improves both capacity and the user
experience even without enhanced interference management schemes introduced in LTE-
Advanced. It can be expected that further increases in the capacity and user experience can
be achieved with LTE-Advanced small cells.

Another study shows a substantial amount of traffic (83%) can be carried on just 10 small
cells in hotspot locations [24]. Similarly, with the use of advanced interference
management schemes, [ 24] shows that 170% increase in cell edge downlink throughput
and 220% increase in median throughput is achieved when 4 small cells are deployed in a
dense macrocell environment (500m macrocell inter-site distance). When 8 small cells
(1732 m macrocell inter-site distance) are deployed, the gains are 170% and 300%
respectively for cell edge and median throughput of users.

Reference [27] compares the user throughput performance gains of small cells using LTE-
Advanced with an interference management scheme in different distribution of users
within the macrocell coverage area[27]. It shows that with 4 small cells per macrocell the
user throughput gain of using small cells can be 10x with user distribution consisting of
hotspot areas compared 2.8x in areas with uniform distribution of users. Similar gains in cell
edge user throughput performance were also reported in [27]. The results suggest that
small cells should be placed very close to the de nse clusters of demand points to maximise
the capacity gain from deploying small cells if no measures are taken to tackle the
interference.

Another study [28] shows that deployment of LTE-Advanced small cells even without
advanced ICIC can contribute to much higher system capacity and user experiences. For
example, it shows that with 10 small cells per macrocell sector, an average throughput of
1117 kbps for macrocell users can be achieved compared with 705 kbps with no small cells
are deployed. Small cell users can enjoy almost 5x gain in user throughput with deployment
of 10 small cells per macrocell and there is also a substantial increase in the system capacity
from 21.2 Mbps to 82.4 Mbps per sector.

These results clearly show how small cells can be used to increase the capacity of a mobile
network. We can also conclude that a higher number of small cells per macrocell can
increase the traffic they relieve from the macrocells and hence increase overall capacity in
the area. Results also suggest that small cells should be placed very close to the demand
points to maximise the capacity gain from deploying small cells if no measures are taken to
tackle the interference.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 37
4.6.2 Radio benefit degradation due to the displacement of the small
cell from the traffic hotspot

To get the maximum radio benefits from the small cell it needs to be co-located with the
demand hotspot. However, it is not always possible to deploy the small cell in the ideal
location (or benefit-maximising location) due to various reasons i.e. deployment is not
practically possible or providing backhaul is either too expensive or not possible. Therefore,
in practice, most of the time the small cell needs to be deployed with a displacement from
the benefit-maximising location.

This displacement will result in the hotspot demand not being fully served by the small cell
or an increase in the number of small cells required to support the hot spot. We model this
effect as an effective reduction in the available capacity of the small cell below its
maximum value as a function of its displacement from the benefit-maximising location. This
function in principle depends on details of the spatial distribution of the demand, the
propagation characteristics of the small cells and the interference and mobility
management processes between the small cell and the macrocell. For modelling simplicity
we have proxied these effects by considering the small cell coverage area and the hotspot
as having circular shapes and calculating the reduction in the overlap area as the centres of
the circles become separated. The result for the parameters in our case study is shown in
Figure 36.

100%
Degradation from maximum benefit

90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0 50 100 150 200 250 300 350
Offset from benefit-maximising location[m]

Figure 36: Benefit degradation vs. placement for case study parameters

4.7 Backhaul options and performance

4.7.1 Backhaul options

A number of backhaul technologies and solutions are available to support small cells
deployments. There is no "universal solution" to fit all requirements and it is expected that

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 38
a mix of solutions will be employed depending on the specific requirements. The
characteristics of differing solutions are described in detail in the Small Cell Forum
document “Backhaul Technologies for Small Cells” [29] and this section summarises some
key points from that reference.

Urban small cell backhaul solutions can be grouped into different categories as follows:

 Carrier frequency range i.e. mm wave, 6 GHz to mm wave range, sub 6 GHz
range
 Propagation type i.e. Line-of-sight (LOS) and non-line-of-sight (NLOS) propagation
 Spectrum licensing arrangement i.e. link licensed, area licensed, light licensed or
licence exempt and dynamic allocation
 Connectivity and topology i.e. point-to-multipoint or point-to-point, forming tree,
ring or mesh

Carrier frequency, propagation type and the operating bandwidth are usually linked to each
other. Lower frequencies, i.e. sub 6 GHz range, possess better wave propagation
characteristics and provide a better penetration through or diffraction around obstructions.
Therefore, lower frequencies are ideally suited to provide NLOS backhaul solutions in the
sub 6 GHz range. However, the amount of spectrum available at lower frequencies is
usually smaller, limiting the operating bandwidth and hence capacity of solutions compared
to those at higher frequencies. However, propagation losses above 6 GHz tend to be high
and usually only LOS propagation is supported at these frequencies. Near Line of Sight
(nLOS) operation is typically used at the low end range of the LOS spectrum. There is no
hard separation between NLOS and LOS solutions as this is highly dependent on the link
budget associated with the system. In the mm wave range, e.g. 60-80 GHz, the absorption
of oxygen in the atmosphere becomes significant in the link budget, thereby limiting the
range.

The use of wireless spectrum is subject to a framework of international and national


regulations. Certain aspects are harmonised at regional level but it is the national regulators
(or government bodies) that define the rules or the technical licensing conditions that the
spectrum users must abide. There are four broad types of licensing categories used for
backhaul spectrum:

 Licence exempt:
o Anyone is free to transmit in this band
o Equipment compliance required
o No fee is payable enabling low cost of operation
o Interference management is complicated
 Link licensed:
o Regulator ensures the interference free operation.
o Co-ordination is required with the regulator
o e.g. traditional point-to-point microwave links
 Light licensed:
o Users are usually shared the information with others during the
coordination process.
o Less expensive than link licensed option.
 Area licensed:
o Licensee is free transmit within the geographic region.
o User is responsible for interference management.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 39
Example allocations of a licensing scheme in the UK, as provided by Ofcom, are shown in
Figure 37. The figure also illustrates three broad ranges of carrier frequencies which are
used to categorise solutions: sub-6 GHz, microwave (6-60 GHz) and millimetre wave (60-80
GHz).

Figure 37: Spectrum allocations for terrestrial services potentially suitable for small cell
backhaul

Table 3 shows a summary of the main wireless backhaul solution categories available [29].
Note that other backhaul solutions, such as satellite links, are mostly used in the rural areas
and not considered as a backhaul solution available in urban areas.

Table 3 Summary of wireless backhaul solution categories

Category Carrier LOS or Non Spectrum Connectivity


name Frequency LOS Licensing

Millimetre
70-80 GHz 71-76/81-86 GHz LOS Light licensed Point-to-point
Millimetre
60 GHz 57-66 GHz LOS Unlicensed Point-to-point
Microwave
point-to-
point 6-56 GHz LOS Link licensed Point-to-point
Microwave
point-to-
multipoint 6-56 GHz LOS Area licensed Point-to-multipoint
Sub 6 GHz 2.4, 3.5 and 5
unlicensed GHz Non LOS Unlicensed Point-to-multipoint
Sub 6 GHz
licensed 800 MHz -6 GHz Non LOS Area licensed Point-to-multipoint

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 40
In this study we have incorporated several types of wireless backhaul solutions. These
solutions are used to provide backhaul to four generic types of small cells and their
associated backhaul links, as shown in Figure 38.

The small cells are divided between those that are served from the macrocell layer down to
the small cells and those that are served from a street-level wireline solution. Wireless
backhaul is optional for these street level solutions.

To other Type A cells if P-MP

Type A Type B

Only if connecting to a Type B

Type D

Type C

Managed wireline dig


NTP

Figure 38: Different urban small cell types and their associated backhaul topology used in
this study

 Type A: These are small cells supported by a wireless connection back to an


urban macrocell. The connection from the macrocell site can be point to point (P-
P) or point to multipoint (P-MP) from an urban macrocell (with an existing high
capacity backhaul link) to a small cell mounted on street furniture or the side of a
building. The small cell can either be integrated within the backhaul unit or can
be a separate device.
 Type B: These are small cells with a wireless backhaul connection which is
provided via the type-A unit backhaul capability. These tend to be from street
furniture to other items of street furniture but could be on the side of a building.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 41
 Type C: These are small cells with a backhaul connection provided to a network
termination point (NTP) by a wireline backhaul solution (this can be either direct
fibre, FTTx, Hybrid fibre or xDSL). Some additional digging may be required to
extend the wireline connectivity to the base of the small cell.
 Type D: These are small cells with a wireless connection which extends the link
from type C to an additional node in an analogous manner that type B extends to
type A.

Hence we can consider that type-A and type-B are ‘macro-down’ solutions and type-C and
type-D are ‘street-level-up’. The number of different types of each small cell required
depends upon the small cell capability and the performance of the backhaul solution. The
proportions of ‘macro down’ solutions (A+B) relative to the ‘street-level up’ in each year of
the study is an input to the model, while the relative proportions of A:B and C:D are
calculated within the model.

4.7.2 Modelling backhaul performance

When a wireless backhaul link applies (types A, B or D) we model its throughput versus
distance performance and check that the link is sustainable with the specified backhaul
technology.

The probability of existence of a line-of-sight varies with distance according to a


geometrical model we have constructed with a form similar to that in [30]. An example of
the form of this model is shown in Figure 39.

1.2

1.0
Line-of-site probability

0.8

0.6

0.4

0.2

0.0
0 100 200 300 400 500 600
Distance [m]

Figure 39: Variation of LOS probability with distance

We then calculate the obstructed and non-obstructed path loss for a given distance using a
free space path loss model. Using link budget parameters and the path loss, the signal to
interference plus noise ratio (SINR) at the receiver end is calculated. We then map the
calculated SINR at the receiver to spectrum efficiency [31] to determine throughput
supported by the link at a given distance using the spectrum efficiency and the bandwidth
of the link for both obstructed and unobstructed cases as shown for one example in Figure

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 42
40. These parameters were calibrated via interviews and data supplied by several wireless
backhaul solution providers.

Non-obstructed throughput (Mbps) Obstructed throughput (Mbps)

160

140

120
Throughput [Mbps]

100

80

60

40

20

0
0 100 200 300 400 500 600
Distance

Figure 40: Variation of throughput versus distance for a particular wireless backhaul
solution

Combining these two allows us to determine the proportion of small cells at a given
distance which can be supported by a given wireless backhaul technology. This is then used
as the basis of the calculation of the relative numbers of each solution type (A:B and C:D) as
described more fully in the next section.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 43
4.7.3 Backhaul dimensioning

Given the LOS probability and the link performance for different wireless backhaul
solutions, we can determine the number of different cells of different types as described
below.

Smallcells

Type D

Type B

Type C
Type A

Managed NTP
wireline Hotspot
Macrocell
Site

Figure 41: Illustration of the deployment of small cells

Capacity congested areas are shown in blue colour polygons covered by small cells whose
backhaul is either anchored at macrocells or at street-level. Single or multiple small cells
will be required to cover the hotspot depending on the size of the hotspot. The hot spot
area and the number of hotspots are inputs to the model. A hotspot is defined as any area
where there is excess traffic that cannot be served by macrocells.

The proportions of “street up” i.e. link types C and D and “macro down” i.e. link types A and
B are assumed to be as shown in Figure 42: the proportion going to street level increases
as the number of small cells increases, reflecting capacity limits on the macro backhaul and
increased street-level fibre availability.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 44
Figure 42: Relative proportion of small cells which use street-up versus macrocell-down
backhaul

Given these inputs the relative number of A:B and C:D cells are determined as described
below.

Number of type-A and type-B cells


Having calculated the total number of small cells required we can also compute the
required backhaul throughputs and the required density of small cells in hotspot areas in
each year of the study. At that throughput the range at which a given backhaul solution can
support the necessary throughput can be determined from the solution performance for
both obstructed and unobstructed conditions as illustrated in Figure 43.

Tput

Unobstructed
Required Tput

Obstructed

Distance
r1 range when r2 range when
obstructed not obstructed

Figure 43: Calculation of obstructed and unobstructed range for a given backhaul
throughput requirement

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 45
Up to range r1, the required throughput can be achieved with either an obstructed or
unobstructed path. All small cells within range r1 can therefore be A-type small cells, as can
some of the cells between ranges r1 and r2 with a probability that depends upon the
probability of being obstructed in that range, as shown in Figure 44 below. This same
principle can be extended to include the obstructed and unobstructed B-type cells that can
link back to A-types.

NLos Bs to LoS As
Los Bs to LoS As

As
Obstructed
As
r r2 r3 D M
As & 1
Obstructed
unobstructed

Bs

Figure 44: Representative coverage areas of A-type and B-type cells showing where line of
sight (unobstructed) and non-LoS (obstructed) links can be secured back to macrocells or
A-type cells.

Assuming uniform traffic within the hotspot and circular hotspot coverage patterns we use
simple geometry to estimate the number of A-type cells out to range r2. We have restricted
this analysis to 2 hops and so all other cells would be B-type cells. Each B-type cell must
connect back to the macrocell through the connection to the A-type cell. This constrains the
number of B-type cells that can connect via the A-type ‘anchor cell’. Iteratively solving the
above system of equations allows us to establish the average number of B-type cells that
can link through each A-type cell and still achieve the required throughput to meet the
small cell traffic demand.

Number of type-C and type-D cells


In general, owing to the higher operating costs of using a leased line, the lifetime cost of a
type-C cell is more than the lifetime cost of a type-D cell. However, if a type-C cell is
deployed with a high capacity wireline backhaul link, it is therefore beneficial to link that
type-C cell to as many type-D cells within the backhaul capacity and latency constraints. In
practice, it is important to consider latency constraints for multihop systems, although this
was not a limiting factor in this analysis. The major cost of installing a type-C link is any
installation cost associated with extending the wireline from the existing point of
termination to the type C cell, either by digging or running cable in an existing containment.
We calculate the average extension distance based on geometrical considerations given the
average density of small cells to be connected and the average density of wireline
termination points in the relevant year. We do this assuming that any deviation from the

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 46
benefit-maximising location is chosen in the direction of the nearest network termination
point, so that it reduces the wireless extension cost accordingly.

Ratio of NC to ND:

We can choose to have as many type-C cells as are required to support the backhaul
requirements of the type-C and connected type-D cells. Similar to the analysis for type-A
and type-B cells, type-D cells will have a maximum distance at which they can successfully
connect back to the type-C cells based on their throughput requirements and probability of
obstructed and unobstructed path. We have conservatively assumed that each type-D cell
cannot be further from the type-C cell than the unobstructed distance to achieve the type-
D cell throughput. This implies that each type-C cell has an effective coverage area equal to
the range of the D-cell backhaul link to provide sufficient throughput plus the range of each
type-D cell. Constraining the maximum number of type-D cells that can link through any
type-C cell, and iteratively solving to maintain the throughput requirements we can
therefore determine the ratio of the number of type-C to type-D cells. Since the number of
type-C plus type-D is an input assumption we can determine the number of type-C and
type-D cells.

4.8 Small cell cost model

For our modelling we have broken the costs of supporting the traffic demand on small cells
to compare to the costs of supporting the same traffic on a macro layer. We consider the
costs in the following three categories:

 Small cell costs


 Backhaul costs
 Core network costs

The cost of each of the elements of the above three categories is described in detail below.

4.8.1 Small cell costs

Below we identify the 2013 costs associated with each small cell and the cost variation over
time. The key elements are:

 Capital cost of each outdoor small cell


 Annual maintenance / licensing cost
 Annual costs associated with planning an integrated Hetnet
 Site acquisition
 Small cell installation
 Utility costs (includes electricity, site rental, business rates)

Small cell capex and opex: Based on the outdoor small cell capex from [32] we have taken
the capital cost for an outdoor small cell to be GBP 4,200, and based on [7] we have taken
opex from maintenance and licensing costs to be 10% of capex. Though small cells are
relatively new to the market we do not anticipate any price erosion in the short term –
instead we anticipate that the functionality included in each small cell (e.g. multi-
technology, multi-band) will increase and the units to be a similar price over the period of

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 47
this study. We have assumed that equipment opex will rise at the average rate of labour
costs of 3.3% (described below).

Site acquisition: Urban small cells need to be sited in appropriate locations. In some cases
the cost per small cell to do this can be reduced if the site used is provided by a landlord
with a large portfolio of sites and negotiations can be performed on a block basis. Real
Wireless estimate that a representative site acquisition cost for an individual site would
require ½ day of effort of a site agent costing GBP350.

Network planning: The need to operate an integrated Hetnet will include additional
network planning effort. However we anticipate that adequate tools and processes will
emerge to allow this task to be performed with negligible impact on operator’s costs and
have not ascribed any cost to this.

Installation: The installation costs per small cell can be reduced by having integrated small
cell/backhaul solutions and by making the connectivity to these as simple as possible. Small
cell suppliers have incorporated this into their system design and small cells can, in general,
be installed simply by installers with no specialist expertise. Installation becomes a matter
of mounting on street furniture or the sides of a buildings and inserting suitable electrical
and backhaul connections. From consultation with stakeholders we estimate that this
would require 1 man-hour of effort costing GBP50 per small cell installation (based on non-
skilled construction labour rates of GBP 20/hour [33] and plant hire for installation). We
have assumed that labour-driven costs will increase at the annual rate of 3.3% from the
same source.

Power: Urban small cells will use the street furniture power supply and pay the supplier on
a bulk basis. An example current product consumes a maximum of 25W [34]. Allowing for
increased power to process multiple technologies we have conservatively assumed that
each small cell will consume 50W continuously, and used the electricity costs for a large
non-domestic user of electricity [35] of 8.93p/kWh with an annual price increase of 6.5%
(midpoint of the low (4%) and high (9%) range).

Site rental: The amount paid for urban cell site rental can be highly variable. Whilst good
sites offer the potential to serve a large number of users efficiently and would demand a
premium, operators are becoming more sensitive to rental costs. Establishing a deal with a
landowner with a large number of available sites would be expected to bring down average
site rental. We have assumed an annual site rental value of GBP 500 per urban small cell
(Real Wireless estimate) which we have assumed will increase with average labour rates.
This rental value would only apply to small cells and not associated backhaul units.

Business rates Cell sites may attract business rates or equivalent base station licensing and
other regulatory costs. Business rates for small cells can be high and industry sources
indicate that they can be a barrier to small cell deployment. We anticipate that business
rates will be reduced by local authorities in order to reduce barriers to deployment and to
facilitate availability of high speed mobile communications. In this study we have assumed
a business rate of GBP 500/small cell, and that this will rise in line with labour costs.

The costs of each of these elements used in this study are summarised in Table 4.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 48
Table 4: Small cell costs used in this report

 Element capex (GBP) capex CAGR opex (GBP) opex CAGR


(2013 values) (%) (for 2013) (%)

Outdoor urban 4200 0 420 3.3


small cell

Urban small cell 50 3.3 0 n/a


installation
Site acquisition 350 3.3 n/a n/a

Network planning n/a 0 0 0


effort per small cell
Electricity n/a n/a 39.01 6.5
Site rental n/a n/a 500 3.3

Business rates n/a n/a 500 3.3

4.8.2 Backhaul costs

Backhaul planning: Types A, B and D require a wireless link. A range of different solutions
spanning frequencies between sub 6GHz up to 80GHz bands and using a range of different
approaches to benefit from propagation effects are used by different solution providers
and is explained in greater detail in section 4.7. Some vendors seek to reduce the
sensitivity of alignment and benefit from MIMO and diversity by using antennas with a
relatively wide main lobe. Others use automatic alignment mechanisms to reduce the
installation effort to install high gain directional antennas. These wireless links require
network planning – but not at the level required to plan RAN networks. We have assumed
that each small cell requires 1 hours of specialist planning effort at a cost of GBP50/small
cell deployed. Different solutions use spectrum that is licensed on a point to point basis,
some use spectrum whose use can be area-based and others use unlicensed spectrum. We
have not set a price for spectrum use for backhaul in this study as we consider this cost is a
minor component of the overall costs and is highly variable by market.

Backhaul controller: Type A backhaul may require installation of a controller at the


macrocell site. Depending upon the vendor solution this controller may link to several small
cells (in the case of a P-MP solution). Some vendors have identical hardware that can act as
a controller or a terminal. Based on consultations we will use a representative cost of a
controller of US$3000 (GBP 1860) and a cost of a terminal of US$1500 (GBP 930). We have
assumed that these prices will remain stable over the period of this study. Annual
maintenance is set at 10% of capex for these units.

Backhaul installation: Sector controllers need to be installed at macrocell sites and this is
more difficult than deploying on street furniture and requires specialist expertise. We have
assumed this will require ½ day of effort and cost GBP 400. Installation of type A terminals
is remote from the macrocell and can be performed with lower skilled technicians.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 49
Assuming 1 person for 1 hour for installation results in an installation cost of GBP 50 per
type A terminal - the same as each small cell above. We have assumed that these units will
re-use backhaul capacity available at the macrocell from an existing backhaul link. We have
assumed that the ‘spare’ capacity available at the macrocell is 500Mbps (half of a 1Gbps
line) and that additional capacity will need to be leased in quanta of 1Gbps when the total
throughput from type A and B small cells linked to macro exceed the available capacity.
Type B and D backhaul solutions require two transceivers. Owing to designing for efficient
deployment, we anticipate that installing both ends of these links can be done in 1 hour
and cost GBP 50. The running costs of type B and D are assumed to be twice the amount of
type A terminals.

Backhaul power: Electricity use and associated costs are assumed to be the same as a small
cell.

Wireline street level backhaul: Type C backhaul solutions use a wireline solution (fibre or
copper-based) to link the small cell mounted on buildings or street furniture to an NTP
(network termination point). Linking the small cell to the NTP can require
groundwork/digging which we have costed at GBP 100/metre dug based on discussions
with industry sources. Electricity demand to interface between the wireline and the
connection to the small cell is assumed to be only 20W. Connection to the wireline is
assumed to take 1 hour and cost GBP 50.

Wireline lease cost: The wireline lease cost is the most significant cost of type C solutions.
The cost of providing fibre connectivity is reducing as more fibre is deployed and industry
migrates to all Ethernet approaches. According to [36] an OC-3 leased line capacity of
155Mbps would cost between $4000 and $7000 per month (GBP 2480 – 4340). However
Ofcom note [37] that BT can provide 1Gbps using Ethernet Access Direct (EAD) on their 21st
century IP core network for GBP 9,500 per year. Based on concerns over significant market
power for the main leased line provider in the UK, Ofcom have capped the price rises that
can be applied to leased lines to be RPI-3.25% [37]. Since RPI is fluctuating around the 2.5%
level over the last 2 years [38] we have assumed leased line rates will reduce by 0.75% /
year over the period of this study. An alternative source is the cost of 100 Mbps enterprise
Ethernet connectivity used in our previous enterprise small cell study [39] at GBP 1,000
capex and GBP 2,000 annual opex.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 50
Table 5: Backhaul capex and opex for the elements needed to support different cell types

 Element capex (GBP) capex CAGR opex (GBP) opex CAGR


(2013 values) (%) (for 2013) (%)

Backhaul network 50 3.3 0 n/a


planning (per small
cell)

Backhaul spectrum fee 0 n/a 0 n/a


(area based fees)

A unit sector 1 860 0 186 3.3


controllers
A unit sector controller 400 3.3 n/a
installation
A units (macro->street) 930 93 3.3
A unit installation 50 3.3 n/a
A unit power 0 39 6.5
A unit spectrum link 0 n/a 0 n/a
fees
B units (macro street- 1 860 0 186 3.3
>street)
B unit installation 50 3.3 n/a
B unit power 0 0 78 6.5
B unit spectrum link 0 0 0 0
fees
C unit cable extension 100 3.3 n/a n/a
(per metre)
C units (cable 50 3.3 n/a n/a
connection)
C unit power n/a n/a 15.65 6.5
C unit cable use/lease 0 10 000 -0.75
D units 1 860 0 186 3.3
D unit install 50 3.3 n/a
D unit power 0 0 78 6.5
D Unit spectrum link 0 0 0 0
fees

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 51
4.8.3 Operational and capital costs of different solution types

An urban small cell (and associated costs) needs to be used with each of the backhaul
solutions identified above. There are differences in the costs and associated equipment
needed for the different small cell types. We can represent the capex (2013 value) and opex
for one year of operation (2013 values) for each of these small cell types as shown in Figure
45.

The totals are summarised in Table 6 below.

Table 6: Summary of costs per small cell type

1000’s GBP Small cell Small cell Small cell Small cell
(2013 values) type A type B type C type D

capex 7.9 6.6 5.7 6.6

First year 1.8 1.7 11.55 1.8


opex
3.56

5 Based on £10k p.a. wireline connectivity and 10m dig length


6 Based on £2k p.a. wireless connectivity and 10m dig length

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 52
Figure 45: capex (2013 values) and opex for one year of operation (2013 values) for each
small cell type

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 53
4.8.4 Core costs

Core functionality required to support operator-controlled small cells will be a marginal


increase upon the existing core functionality. We have allowed a small cell gateway whose
primary purpose will be to aggregate small cell traffic to reduce the number of separate
links needed to support the traffic carried on small cells [40]. The study area considered in
this report is based on a small metropolitan area – in practice a gateway would support a
large number of small cells, though its costs would be ‘up front’. We have determined the
contribution of the cost for each small cell and attributed capex in the year that additional
small cells are deployed.

We have assumed that this cost will be no more than the cost of a local controller used in
the enterprise small cells and would cost $40 (GBP 24.80) per small cell supported and used
10% of the capex as the opex. Equipment prices are assumed to be stable and operating
costs will rise with the average labour rate as described above. Outcomes are shown in
Table 7.

Table 7: Core network capex and opex (GBP k) for the elements needed to support small
cells

 Element capex capex CAGR opex (GBP) opex CAGR


(GBP) (%) (for 2013) (%)
(2013
values)

Small cell gateway 24.80 0 2.48 3.3


(cost per small cell
supported)

4.8.5 Summary of costs and comparison with macrocells

Based on the above figure and using a discount value of 8%, we can determine the costs of
ownership of each individual small cell type over its lifetime (2014 to 2020) as shown in
Table 8. For comparison, Ofcom’s LRIC model estimates the capex and opex of a generic
macrocell to be GBP 134,000 and GBP 15,000 respectively [41] which corresponds to a PV
of GBP 217k if we use a 6 year period and a discount rate of 8% as we have assumed for the
small cells. Hence for type A, B and D backhaul the TCO is within 10% of a macrocell cost
and wireline capex dominates the cost of Type C cells.

Table 8: Total Cost of Ownership for different cell types

Figure of merit Type A Type B Type C Type D

TCO 19.8 18.0 84.27 18.0


(000’s GBP, 2013 38.08
value)

7 Based on £10k p.a. wireline backhaul opex and 100m dig length
8 Based on £2k p.a. wireline backhaul opex and 100m dig length

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 54
(see below)

4.9 Approach to financial modelling

An operator may choose to deploy small cells where it will add value to the network. Whilst
we note that adding small cells to the network can have many positive impacts on users
experience in this financial assessment we have taken the restrictive view of assessing the
financial benefits to small cells as a matter of comparing the cost of delivering traffic using
small cells or using macrocells.

The cost elements are based on supporting delivery of the traffic demand over small cells
using the cost elements from section 4.8 and the assumed mix of backhaul solutions in
section 4.7.3. The ‘savings’ are simply the costs that would have been incurred had the
same traffic been delivered on a macrocell layers using the cost per GByte from section
4.4.1.

The following key financial metrics are output from the model:

 Total Cost of Ownership (TCO) – this is calculated as the total capex and opex of
the solution over the system lifetime (assumed to be 7 years 2014-2020) without
discounts
 Net Present Value (NPV) – this is the 2013 value of the system taking the costs
away from the benefits of the lifetime of the system and discounting to 2013
values.
 Return on Investment (RoI) – this is the proportion of the benefits of operating
the system (profits – costs) compared to the cost of investment (all profits and
costs discounted to 2013 values)
 Payback Period (PP) – this is the time taken for the cumulative benefits to exceed
the cumulative costs (with no discounting)

All income and expenditure is calculated as at the end of the year in which they occur.

We have taken the lifetime of the small cell network to be 7 years (beginning of 2014 to the
end of 2020). Small cell equipment is assumed to have a lifetime of 5 years and any
installed equipment is replaced every 5 years. After 5 years the network will still be in place
and able to be a source of additional value. In order to assess the value (so-called terminal
value) of the small cell network after the 7 year period we:

 Freeze the macrocell cost / GB at the value attained at the end of 2020
 Maintain the same traffic level (i.e. no growth) on the small cell network at the
2020 level
 Continue to follow the 5 year replacement period cycle
 Project the incomes and expenditures for the next 10 years and discount the
associated net cashflow to 2013 values. This is the terminal value determined in
this analysis.

We additionally calculate the extended RoI (i.e. the RoI calculated over the period 2014-
2030) assuming the network is operated over the period 2020-2030 and using the
assumptions to determine the terminal value.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 55
4.10 Capacity-driven deployment case study

4.10.1 Capacity-limited scenario description

The parameters for this scenario have been mainly described in preceding sections, but are
summarised together in Table 9.

Table 9: Base case scenario parameters

Parameter Initial Final Units


Value value
(2014) (2020)

Starting cost of carrying 1 GB in 0.2 Linked to scalar


macrocell sites demand
density

Cost/GB rise index 1.0 GBP /GB


Proportion of macro down (type 0.9 0.5
A + type B) links

Number of macrocell sites 130 136

% of traffic offloaded to indoor 77 90 %

Spectrum efficiency of macrocell 2.16 3.7 bps/Hz


sites

Spectrum efficiency of small cell 0.9 1.53 bps/Hz


sites

Total spectrum used by 40 40 MHz


macrocell sites

Total spectrum used by small cell 20 40 MHz


sites

Wireline cabinet density 330 330 Cabinets / km2


Demand growth rate variable 35 %

Total demand in study area 81000 TB/month

4.10.2 Solution description

Using the methodology identified above the number of small cells grew from 93 needed in
2014 to a total of 1119 in 2020 as shown in Figure 46. This results in some nine small cells
per macrocell in the study area at the end of the period.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 56
Number of small cells
Number of small cells required Average number of small cell per macro

1200 10
9
1000
8

Av number of smallcells per macrocell site


Number of smallcells

7
800
6
600 5
4
400
3
2
200
1
0 0
2014 2015 2016 2017 2018 2019 2020
Year

Figure 46: Calculated number of small cells needed

The computed proportion of the different backhaul link types is shown in Figure 47. Note
that the ratio of D to C type cells increased to the maximum permitted value of 5 in 2015.

Number of Ds Number of Cs Number of Bs Number of As

1200

1000
Numberof links

800

600

400

200

0
2014 2015 2016 2017 2018 2019 2020
Year

Figure 47: Number of small cells of each type required to support the baseline case

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 57
The ratio of A/(A+B) grew throughput the period reflecting the difficulty of accommodating
the combined backhaul of connected type-B and type-A cells through the same type-A
backhaul link.

4.10.3 Outcomes – base case results

Based on the above modelling parameters, the cumulative costs, saving and cash flows are
shown in Figure 48.

Cumulative costs, savings and cash flow


80000000
Cost, savings and cashflow (GBP)

70000000
60000000
50000000
40000000
30000000
20000000
10000000
0
2014 2015 2016 2017 2018 2019 2020
-10000000

Cumulative Costs Cumulative Savings Cumulative Cash Flow

Figure 48: Baseline cumulative costs, savings and net cash flows

Using the discount rate of 8%, the other financial results are shown in Table 10.

Table 10: Key financial outcomes for base case

Financial Metric Value


TCO £18,466,533
NPV (study period) £17,364,926
NPV (including terminal value) £114,982,761
ROI 136%
Payback Period 3.78 years
SC Network Terminal Value £97,617,836
Terminal value / NPV 5.62

The total cost of ownership to support the traffic is GBP 18.5M, with an NPV of nearly GBP
17.4M. The terminal value of the small cell network is nearly 6 times the NPV over the 7
year study period reflecting a high return on ongoing small cell investment growth once the
initial cost of establishing the small cell layer has been recouped. Hence there is a strong
business case for this investment.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 58
4.10.4 Outcomes – sensitivity analysis

In order to understand the sensitivity of the base case results of the assumed parameters
we have varied different parameter values in order to investigate the impact that these
have on the NPV during the study period. The parameters varied are:

 The total traffic demand growth rate (CAGR)


 The cost of carrying traffic on the macrocell layer
 The number of macrocells deployed in the study area
 The percentage of traffic offloaded to indoor cells
 The macrocell and small cell spectrum efficiencies
 The spectrum used in both macrocells and small cells
 The offset of the small cell location from the benefit-maximising location (BML)
 Wireline cabinet density in the study area

For some attributes the parameter value is allowed to vary throughout the study period,
and where this is the case the initial and final value are shown on the x-axis. Owing to the
large terminal value compared to the NPV over a short 7 year study period we have
separately identified both the NPV over the study period and the terminal value of the
small cell network.

These sensitivities are shown in graphical form as shown in Figure 49 to

Wireline cabinet density


£140,000,000
£120,000,000
£100,000,000
£80,000,000
NPV

£60,000,000 Terminal Value


£40,000,000
NPV (study period)
£20,000,000
£0
200 330 400
Wireline NTE density (NTEs/km2)

Figure 58 and discussed below. The base case values are coloured green.

Traffic demand CAGR


£400,000,000
£350,000,000
£300,000,000 NPV
£250,000,000 (study
NPV

£200,000,000 period)
£150,000,000
£100,000,000 NPV (incl
£50,000,000 terminal
£0 value)
20% 25% 30% 35% 40% 45% 50%
Rate of demand increase

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 59
Figure 49: Impact of different traffic demand growth rates on net present value

Macrocell cost /GB parameters


£600,000,000
£500,000,000
£400,000,000
NPV

£300,000,000
£200,000,000 Terminal Value
£100,000,000 NPV (study period)
£0
0.41, 0 0.15, 1 0.2, 1 0.25, 1 0.15, 2
Cost per GB (2013 base level, growth index value)

Figure 50: Impact of macrocell cost on the NPV

Macrocell density
£150,000,000

£100,000,000
NPV

£50,000,000 Terminal value

£0 NPV (study period)


120 - 120 130 - 136 160 - 160
Number of macros in study area (initial, final)

Figure 51: Impact of the number of macrocells on the NPV

% traffic offloaded to indoor


£180,000,000
£160,000,000
£140,000,000
£120,000,000
£100,000,000
NPV

£80,000,000 Terminal Value


£60,000,000
£40,000,000 NPV (study period)
£20,000,000
£0
77 - 77 77 - 90 90 - 90
% traffic offloaded to indoor (initial, final)

Figure 52: Impact of different traffic offload on the NPV

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 60
Spectrum efficiency of macrocell
£160,000,000
£140,000,000
£120,000,000
£100,000,000

NPV
£80,000,000
Terminal Value
£60,000,000
£40,000,000 NPV (study period)
£20,000,000
£0
2.16 - 2.16 2.16 - 3.7 3.7 - 3.7
Spectrum efficiency (bps/Hz) (initial, final)

Figure 53: Impact of macrocell spectrum efficiency on the NPV

Spectrum efficiency of small cell


£140,000,000
£120,000,000
£100,000,000
£80,000,000
NPV

£60,000,000 Terminal Value


£40,000,000 NPV (study period)
£20,000,000
£0
0.89 - 0.89 0.89 - 1.53 1.53 - 1.53
Spectrum efficiency (bps/Hz) (initial, final)

Figure 54: Impact of small cell spectrum efficiency on the NPV

Total spectrum used by macrocell


£140,000,000
£120,000,000
£100,000,000
£80,000,000
NPV

£60,000,000 Terminal Value


£40,000,000 NPV (study period)
£20,000,000
£0
30-30 40-40 50-50
Downlink spectrum available on macrocell layer (initial, final)

Figure 55: Impact of spectrum available on the macrocell layer on the NPV

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 61
Total spectrum used by small cells
£140,000,000

£120,000,000

£100,000,000

£80,000,000
NPV

£60,000,000

£40,000,000

£20,000,000

£0
10-30 20-40 30-40
Downlink spectrum available on small cells (initial, final)

Figure 56: Impact of spectrum available on small cells

Offset from BML


£140,000,000
£120,000,000 NPV
£100,000,000 (study
£80,000,000 period)
NPV

£60,000,000 NPV (incl


£40,000,000 terminal
£20,000,000 value)
£0
0 5 10 15 20 25
Offset from BML (m)

Figure 57: Impact of changes on the Benefit-maximising Location on the NPV

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 62
Wireline cabinet density
£140,000,000
£120,000,000
£100,000,000
£80,000,000

NPV
£60,000,000 Terminal Value
£40,000,000
NPV (study period)
£20,000,000
£0
200 330 400
Wireline NTE density (NTEs/km2)

Figure 58: Impact of changes on wireline NTE cabinet density on the NPV

All of the following comments are based on variations around the baseline case and
changes identified may not apply in other circumstances.

Traffic growth: When traffic demand grows at less than 12% annually, anticipated spectrum
efficiency improvements in the macrocell layer are able to absorb all additional traffic
beyond 2015 and so there is no positive case for urban small cells. However most markets
expect growth substantially beyond this level and the business case improves continuously
as demand growth increases.

Cost of incremental capacity on the macrocell layer: If the macrocell layer cost is held
constant over the study period the NPV becomes positive whenever the cost of incremental
capacity on the macrocell layer exceeds GBP 0.4/GB. However, we expect the cost to rise in
relation to the relevant demand density. The baseline case used GBP 0.2/GB growing in
proportion to the increased outdoor traffic density. The price increase per GB in capacity
constrained environments is highly dependent on an individual operator’s network and
spectrum assets.

The number of macrocells deployed in the study area: An increased number of macrocells
are able to support more traffic (which would leave less traffic for the small cell layer)
resulting in a reduced NPV for small cells over the time period of the study.

The percentage of traffic carried by indoor cells: Varying the traffic offloaded to indoor
cells changes the traffic to be carried by the macro and small cell layers. Counterintuitively,
offloading 77% initially (2014) and 90% (2020) is less positive than offloading either 77% or
90% throughout. This is explained by the timing of the need to deploy small cells and the
ability to utilise them fully at a later date when the cost per GB is assumed to be higher.
Offloading 77% results in a larger amount of traffic to be served in the longer term.
Offloading 90% results in deferring purchase of small cells (that would also need to be
replaced in the study time period) until the cost per GB has risen and the value of deploying
small cells is greater.

The macrocell and small cell spectrum efficiencies: A failure to improve macrocell
efficiencies results in more small cells being deployed and leads to a higher NPV. Having a
high macrocell spectrum efficiency delays small cell deployment until the cost per GB has

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 63
increased. Any improvement in small cell spectrum efficiency is positive since less cells and
consequently lower expenditure is required to support the same traffic.

The spectrum used in both macrocells and small cells: Less spectrum in the macrocell layer
(30MHz) results in more traffic to be carried by small cells. More spectrum (50MHz) results
in deferral of small cell deployment until the assumed cost per GB is more but leads to less
traffic to be supported on the small cells. Similarly the more spectrum that can be deployed
in the small cell layer results in more traffic to be supported by less cells and leads to a
more positive business case. The terminal values of the resulting small cells are more when
less spectrum is deployed in the macrocell layer.

The offset from the benefit-maximising location: Any change from the BML is assumed to
reduce the cost of any digging required but results in less efficient targeting of the traffic.
For the baseline values these two conflicting trade-offs result in an optimum trade-off
approximately 5m from the position that would maximise the traffic served per small cell.
With the baseline cabinet densities of 330 cabinets/km2 this optimum position is not
sensitive to this offset distance.

Wireline cabinet density variability: The NPV does not vary significantly with reasonably
large changes in the assumed wireline NTE density. Neglecting any offset from the BML, the
average dig length varies inversely as the square root of the average cabinet density and at
400 cabinets/km2 has an average dig length9 of 17.5m. This one-off cost is not a strong
driver compared to the assumed wireline opex costs.

4.10.5 Findings

Based on the analysis used to derive the model the baseline results and the sensitivity
analysis identified above we can determine:

 Timing is critical for the success of marginal business cases for some small cell
deployments. In particular small cells maximise their benefit when they can take
a large amount of traffic so operators should aim to deploy their small cell assets
as close as possible to the time at which they become more cost effective than
the macrocell layer.
 Once the small cell layer has been deployed the initial costs of establishing the
layer are sunk and subsequent investment in small cells becomes very attractive.
This suggests that upgrades of small cells to increase their capacity via greater
spectrum or spectrum efficiency will be highly cost-effective.
 There is an optimal trade-off in locating the small cell exactly in the position of
the traffic and close to available backhaul. Given the assumed baseline
parameters, the optimal location is not highly sensitive to small deviations from
this optimal position. The sensitivity will depend upon the dig costs and their
proportion of the TCO of cells that use wireline solutions.
With the baseline parameters the dig costs and installation costs are a relatively
small cost element of the TCO of a wireline backhaul solution but the leased line
costs are a large cost element.

9Assuming uniform distribution of cabinets and small cells, integrating over each quadrant of the resulting grid,
the average distance to a cabinet is 500*0.7/(density0.5) m.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 64
Overall it is clear that small cells play a vital role in providing a cost-effective means of
enhancing capacity once demand density exceeds a critical level. This critical level will
depend significantly on the operator’s existing network and spectrum assets which need to
be determined on an individual basis. Once small cells have been deployed initially they
form a very cost-effective approach to subsequent capacity enhancement.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 65
5. City centre small cell business case – coverage-driven
deployment

5.1 Scenario description

The analysis in the previous chapter established a business case for urban small cells driven
by a need for capacity enhancement. In this chapter we look at an alternative market
driver, that of the need for enhancement of coverage depth. While city centres are well
covered by macrocells in terms of the overall area, the depth and consistency of coverage
may not always be adequate. In some cases these deficiencies may increase with time due
to changing building construction materials and user expectation of higher data rates to
constitute an acceptable service. The challenges may be most acute for an operator whose
spectrum portfolio does not include low frequency (e.g. below 1 GHz) spectrum.

This is the case we examine here: an operator has an existing macrocell network in a dense
urban area providing good extent of outdoor coverage but has no sub-1GHz spectrum.
While their macrocell network is as dense as their competitors, their network has limited
coverage depth. The operator decides to deploy a layer of outdoor small cells to deliver
coverage depth and user experience at or beyond that of its competitors.

In the absence of small cells the operator would need to increase the macrocell density to
achieve target coverage depth given specified building penetration loss and path loss
assumptions. It would then experience additional costs relative to its competitors in
building and operating additional macrocells. We seek to determine whether the small cell
costs can be less than the costs of achieving the same outcome via macrocells, and also to
determine what additional benefits a small cell layer may provide.

5.2 Approach to calculating radio parameters

To evaluate the coverage driven scenario, we assumed both operators have same macro
cell site density i.e. 3.65 macro cell sites per km2. We developed a link budget consistent
with [42] to obtain the maximum allowable path loss for LTE macrocell networks deployed
in the 2.6 GHz and 800 MHz frequency bands and for small cells deployed at 2.6 GHz. Using
the link budget parameters we calculate the maximum cell range for a given throughput i.e.
service level and depth of coverage using path loss models. The path loss model provided in
[43] is used to calculate the maximum cell range for macrocells. For small cells, the
maximum cell range is calculated using the path loss model provided in [10].

We analysed the following cases between two operators:

1. Difference in service levels if both operators are providing services at the same
depth of coverage.
2. Difference in depth of coverage if the same level of services are provided by both
operators.
One option available to operator A, to match the services provided by operator B, is to
increase the number of macrocell sites. We provide the number of macro cell sites required
to match the service level of the 800 MHz operator. Alternatively operator A can deploy
small cells to provide the same level of services provide by the operator B. In that case
operator A can either deploy small cells in the:

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 66
 whole area to provide significant service level benefits to its users compared to
operator B
 targetted in the area where operator A’s macrocell services are not available or
below expected levels to match the service levels provided by operator B.

Both these cases were examined and used to drive the financial outcomes described below.

5.3 Approach to financial modelling

The financial benefits of deploying small cells stem from two key aspects:

 Cost benefit of reduced cost of coverage for a given quality level using small cells
instead of additional macrocells
 Benefit by having improved quality of service (improved data rates and improved
depth of coverage compared to that able to be provided by the macrocell layer)

The small cell unit costs previously used for the capacity case have been re-used in this case
study. Here, however, we are comparing the cost of providing service using small cells to
that of providing service using macrocells. Since macrocells have a relatively long
equipment lifetime, we have derived the 20 year present value cost, using a 5 year
equipment replacement cycle for the small cells. We have neglected any terminal value for
either small cells or macrocells. The same 8% discount rate used previously has been
applied.

The cost of coverage using macrocells depends upon the complexity of establishing suitable
sites in these difficult to cover areas. Civil works and site rental values have a large
variation in these circumstances. We can split the difficulty of establishing these macrocells
into low, medium and high cost deployments. In a previous study [44] the capex cost of
deploying rural macro sites ranged from GBP 88k to GBP 253k and had an opex of GBP 16k.
Opex, particularly rental values, in dense urban areas would tend to increase above these
values, and equipment costs have reduced. Ofcom’s 2014 LRIC model has capex of GBP
134k and GBP 15k opex. For the purposes of this study, we have used the values shown in
Table 11, which reflect increased opex, and reduced equipment costs from these previous
studies and are consistent with the aforementioned Ericsson study. The site acquisition and
civil works are high and reflect our understanding of the difficulty in deploying additional
macrocell sites in dense urban areas.

The ratio of these different macrocell types will depend upon the environment to be
covered. For the purposes of this study we have used the proportions of low, medium and
high cost macrocells as shown in Table 12, resulting in the average macrocell costs shown in
Table 11.

Table 11: Cost elements for low, medium and high cost macrocells (GBP, 2013 values)

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 67
Cost element Low Medium High Weighted
Average

Site Acquisition
and civil works 31800 71450 204200 56970

Equipment
costs 26900 26900 26900 26900

Operating costs
(2013) 17650 19600 22200 18495

Table 12: Proportion of low, medium and high cost macrocells

Low Medium High

0.7 0.2 0.1

5.3.1 Deriving additional benefit

The additional benefit resulting from improved coverage and data rates can be difficult to
quantify. Many different sources of benefit can be attributed to an improved mobile
service, including:

 Improvements in business productivity for mobile workers: Improved voice


coverage will allow workers to maintain contact with their colleagues, suppliers
and customers and use their time more efficiently whilst away from their work
base. Improved mobile speeds will allow new mobile applications to be used to
improve access to company data and /or to improve process automation. There
will be a lag in this benefit owing to the need to change procedures.
 Improved public sector efficiency: Public sector workers will benefit in a similar
manner to the above (e.g. police officers could spend less time on paperwork and
gain access to decision making information in a more timely manner). We do not
anticipate increased spending by the public sector. The economic value will be in
the quality of service delivery rather than in changing the economic output
associated with the public sector.

There are other economic benefits of improved mobile services, such as:
 Infrastructure spending for additional network elements: We will not include this
as we are seeking to identify the benefits of the use of mobile connectivity
 New business opportunities: Availability of mobile communications has changed
how some businesses function. Any opportunities are likely to develop owing to
the existing network availability and it would be difficult to assign an economic
value to the incremental improvements anticipated in this study.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 68
 Increased consumer surplus: Consumers are likely to derive increased benefit
from a service that allows improved connectivity in more environments. This
benefit is enjoyed by consumers but does not contribute to GDP.

For the purposes of this study we will identify the additional economic benefit that can be
derived from improvements in private sector business productivity and note that other
benefits, though less tangible, will also exist.

According to the UK Broadband Impact Study [45], a doubling of in-year data-rates for fixed
broadband results in an uplift in productivity of 0.3%. Ofcom’s Communication Market
Report 2013, notes that in 2012 average mobile phone use was roughly half that of fixed
internet use (29mins versus 68mins). Ascribing no additional benefit to mobile use than
fixed, we can estimate the uplift in an in-year doubling of data rate for mobile phones to be
29/68×0.3% = 0.13%. It is assumed that the scale of any benefits will diminish with
additional in-year speed improvements and we have assumed that any benefit would not
exceed double the benefit of a 100% speed increase. The relationship we have derived to
have these attributes is shown in Figure 59.

Productivity improvement versus in-year datarate


speed increase
0.25
Productivity uplift (%)

0.2
0.15
0.1
0.05
0
0% 50% 100% 150% 200% 250% 300% 350%
In-year datarate speed Improvement (%)

Figure 59: Assumed relationship between productivity uplift and in-year mobile data
speed improvement factor

Daytime population rise: Dense urban areas, like the one in this case study, typically have
larger population densities during the daytime busy hour than are resident in the area. The
London Data Centre [46] identifies population statistics, including daytime and registered
resident numbers and the number of employees. The ratio between daytime and evening
population ranging between boroughs with a peak of 80 for the City of London, 5 for
Westminster approximate parity for less dense urban areas. Representative values for
dense urban areas ignoring these peaks is approximately 1.5, and 80% of the daytime
population are employed. According to the Office of National Statistics [47] 5.9m people
are employed in the public sector in the UK (19% of the workforce). We will therefore
assume that 80% of the daytime population who work are employed in the private sector.

The Gross Value Added per worker: This is defined as the value of goods and services
produced by each worker in the area of their work. In the UK, the Office of National
Statistics publishes GVA in the UK [48], which varies significantly according to the nature of
the work done in different regions, ranging from GBP20K to GBP36K in urban areas, with a
national average of GBP20.9K and London being the most productive area with a GVA per

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 69
head of GBP35.6K. For the purposes of this study we will assume that the GVA in the dense
urban environment is GBP25K per daytime worker in the area.

Benefit to the Operator: Whenever data connectivity and/or data transfer rates improve
subscribers derive benefit from the improved connectivity and will improve productivity as
described above. Operators would seek to gain additional income for being able to provide
higher levels of service than competitors. Research from the UK government [49] finds that
user benefits of mobile communications are typically 6 times higher than the operator
benefits. We will therefore conservatively assume that operators can extract a value
equivalent to 1/6 of the increased productivity benefit.

5.3.2 The cost of 700/800MHz spectrum

This case study compares the ability of an operator with only high frequency spectrum to
establish coverage and target data rates with small cells that are achievable with an
operator who has a macro layer using sub-1GHz spectrum.

DotEcon and Aetha performed an assessment of spectrum value for Ofcom [50] and found
that average prices in Europe were in the range 0.13 to 0.71 GBP/MHz/pop with an average
of 0.47GBP/MHz/pop.

The prices paid for 700MHz spectrum in the US in 2008 are summarised at [51]. Though
some of the blocks were regional a total of 62MHz was released and raised slightly over
USD19Bn. Taking the US population as just over 316million yields a spectrum valuation of
USD0.97/MHz/pop (GBP 0.59/MHz/pop) – close to the European average. We will use a
spectrum valuation of 1USD/MHz/pop in the study area to assess any benefit in using small
cells instead of procuring sub-1GHz spectrum.

5.4 Outcomes

For the baseline case we derived the performance available for operator A (assumed to
only have access to 2.6GHz spectrum) and operator B (with 2x10MHz of sub 1GHz
spectrum), assuming that operator A attempts to match operator B for either indoor
coverage depth or edge of cell data rate. Operator A is unable to compete effectively in
terms of data speed or coverage depth as shown in Table 13 and Table 14.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 70
Table 13: Coverage performance comparison when equal depth of coverage sought

Operator A Operator B

Criteria Coverage Tput Coverage Tput


depth (m) (Mbps) depth (m) (Mbps)

Depth coverage
target 15 15
Tput achieved 21.1 32.24

Table 14: Coverage performance comparison when equal data rate sought

Operator A Operator B

Criteria Coverage Tput Coverage Tput


depth (m) (Mbps) depth (m) (Mbps)

Tput datarate
target 35.17 35.17
Indoor depth
achieved 0.5 10
In the following we present the cells required, the costs and additional benefits of operator
A augmenting the existing network in order to match or exceed the performance attainable
by operator B. Operator A is assumed to have three routes to achieve this:

 Augment the existing macrocell layer with more macrocells


 Widespread deployment of small cells to cover the study area
 Targeted deployment of small cells in the areas where target performance with
the existing macro layer does not exceed that of operator B.

5.4.1 Augment existing macrocell layer

The number of cells and their cost over the 20 year macrocell lifetime (2013 values) are
shown in Table 15.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 71
Table 15: Number of macrocells and associated costs

Requirement Match Tput Match Depth


of coverage

Additional
Macrocells 79 60
PV Cost (2013)
($m) 39.4 30.0

5.4.2 Deployment of small cells across the entire study area

We anticipate that small cells will provide an additional benefits than coverage owing to
improvements in geometry factors and associated data rate increases. The number of small
cells, costs and associated benefits are shown in Table 16.

Table 16: Number of small cells with associated costs and benefits

Requirement Match Tput Match Depth


of coverage

Small cells required 1387 1510

PV Cost 2013 ($m) 85.6 93.2


Worker efficiency gain
($m) 16.6 16.7
Assumed added value to
operator ($m) 2.8 2.8

Clearly the additional benefits to the operator do not compensate for the additional cost of
providing these small cells, and this is unlikely to be an attractive deployment option.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 72
5.4.3 Targeted deployment of small cells

In this scenario small cells are targeted at the areas where the existing macrocell layer
performance is not adequate. The number of small cells, costs and associated benefits are
shown in

Table 17: Number of small cells with associated costs and benefits

Requirement Match Tput Match Depth


of coverage

Small cells required 440 523

PV Cost 2013 ($m) 27.2 32.3


Worker efficiency gain
($m) 11.3 6.3
Assumed added value to
operator ($m) 1.8 1.0
Spectrum purchase
savings 5.0 5.0

For these different performance criteria, targeted deployment of small cells is more cost
effective than augmenting the macro layer and offers several additional advantages. These
advantages are only quantified for the part of the study area where small cells are used to
improve the network performance – though the performance of the existing macro layer
would also improve (though is not quantified here).

5.5 Findings

As is well understood, the use of sub 1GHz spectrum can allow improved indoor coverage
compared to higher frequency spectrum. A key issue for operators who do not have sub-
1GHz spectrum is how to improve the quality and depth of service as users expect
ubiquitous high data rate availability. We have found that:

 Augmenting the macrocell layer with additional macrocells can match the service
achievable by an operator with sub 1GHz spectrum but at significant cost.
 Widespread coverage of the dense urban area with small cells can also match the
service – but is more costly than augmenting coverage with macrocells
 Targeted deployment of the inadequately served parts of the study area with
small cells is less expensive than using macrocells and provides additional
advantages. These additional advantages bring efficiency benefits to workers
ranging from $6.3m to $11.3m and provide the operator the opportunity to
benefit by between $1.0m and $1.9m. These advantages can be realised without
investment in sub-1GHz spectrum, which would represent another saving of $5m
in the coverage area at contemporary values of spectrum revealed at auctions.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 73
6. Transport hub small cell business case

6.1 Background

Urban small deployments include not only outdoor city centre situations, but a wide range
of indoor environments where public access is possible. As described earlier these include
retail, entertainment and transportation environments.

Urban small cell deployments in indoor locations are similar in some ways to the enterprise
deployments we described in our paper on the business case for enterprise small cells [2].
They will involve high capacity access points interlinked via a LAN and managed by the
operator. However, differences include:

 Public, open access


 Potentially far higher capacity density requirements
 Transient population
 Potential for location-specific retail and entertainment applications

Operators see small cells as playing an important role, as exemplified by the following
quotation:

“Mass gatherings, such as sports stadiums, need special solutions. Last year the UK telecoms industry
performed very well throughout the Olympics 2012, having prepared well in advance and deployed lots
of capacity.”

- Andy Sutton, Principal Network Architect at EE UK on their future small cell plans [52]

In order to explore the business case for such indoor urban small cells, we have selected a
transport hub as a case study – specifically, a busy city-centre railway station.

6.2 Case study description

Our case study examines the needs of an operator who is faced with very high demand at
peak times in a large railway terminus. The parameters of the station are similar to those of
King’s Cross railway station in London10.

10
Some readers may be interested to know that this is the same station featured in the Harry Potter series.
Note however that we have not included any traffic from passengers on platform 9¾ in our analysis.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 74
Figure 60: King's Cross Railway Station

The parameters are drawn from a previous study conducted by Real Wireless regarding the
capacity of 4G systems [53]. This study gives the station as carrying some 73,000
passengers during the morning peak (7a.m. -10a.m.) in an accessible area of 8,000m2.

If we assume 20 minutes dwell time in the station for each passenger, this yields an average
population density of 1,014,000 people /km2. This is considerably in excess of the density in
the wider city even within the urban core, as previously illustrated in section 2.1, being over
30x the dense urban population density. Taking into account other factors (proportion of
high data users and indoor/outdoor mix) we estimate x27.6 traffic density relative to a
dense urban baseline. Given that this puts the passengers into what anthropologists refer
to as intimate space (see Figure 61) this is amongst the highest densities of users which
operators will ever encounter. Those users are also likely to be very active on their mobile
devices at peak times, checking train times, sending messages and consuming news and
other video content.

Figure 61: Categorisation of ranges of user density

This type of hyperdense traffic requires special solutions. In the past operators have used
macrocell-type equipment with distributed antenna systems in such locations, but when
such systems are sectorised intensely they reach a point where a single antenna per sector
is provided and the distributed system serves no benefit. Denser reuse is then required,
suggesting a small cell architecture.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 75
The challenge is to determine whether such an architecture can be cost-effective in
delivering capacity as demand continues to rise.

6.3 Solution description

We have estimated the incremental cost of capacity for a small cell network deployed in
the railway station to meet the demand from one operator’s customers in the station. The
operator has a 25% market share and initially deploys LTE small cells in 2014, dimensioned
to meet the entirety of demand from its users. We assume that the number of passengers
stays constant over the study period to 2020. Each year new small cells are built to
accommodate incremental demand, with both the capital costs of deploying the small cells
and the operating and replacement costs incurred over the whole period.

Other assumptions and parameters are shown in the following table:

Table 18: Parameters and assumptions for railway station case study

Parameter Value Unit Comment


Operator market share 25% % Percentage of users with the one operator
Users in station during 8,111 Mobile users Based on known footfall of station, 8,000
busy hours m2 accessible area, 3 hours busy period and
20 minute dwell time
Traffic growth per user 35% % Compound annual growth rate during study
period, as for the base case in our capacity-
drive study. The demand per user is also
consistent with that study with a 10%
activity factor.
Small cell spectrum 1.5 bps/Hz Assumption based on enhanced geometry
efficiency factor for indoor small cells
Spectrum used by small 2 x 20 MHz Assumed to be in a higher frequency band
cells such as 1.8 GHz or 2.6GHz

Based on these values we compute the number of small cells required to deliver the
necessary capacity density. Even at the start of the period the system is capacity limited. As
the demand grows, so the small cell density is increased to cope with the demand, as
shown in Figure 62.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 76
200

Number of small cells to serve capacity


180
160
140

in railway station
120
100
80
60
40
20
-
- 0.10 0.20 0.30 0.40 0.50 0.60 0.70
Traffic density in busy hour (Mbps/m2)

Figure 62: Small cells grow in proportion to demand in capacity-limited environment

The costs of the small cells are modelled similarly to those presented in our work on
enterprise small cells, based on the highest capacity variant in that report: see section 7.7 in
[2] for details of the unit costs applied and the methodology used for attributing these costs
over time, including both capex and opex from building, replacing and operating small cells
and the supporting backhaul. The associated costs incurred each year are determined
according to the models in that reference and shown in Figure 63.

250,000
Annual expenditure on building and

200,000
operating small cells

150,000

100,000

50,000

0
2014 2015 2016 2017 2018 2019 2020

Capex + replacement costs £ Opex £

Figure 63: Expenditure associated with initial build and subsequent expansion of small
cells in railway station

We then calculate these costs on the basis of present values in 2013 and determine the
incremental cost per gigabyte of carrying this traffic, resulting in the reducing cost per GB
shown in Figure 64. The diminishing cost arises from the initial investment in small cells in
the first year.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 77
0.025 0.016

0.014

Incremental cost per GB of carrying


0.020

traffic on small cell network


0.012

0.010
0.015

(£/GB)

($/GB)
0.008

0.010
0.006

0.004
0.005
0.002

0.000 0.000
2013 2014 2015 2016 2017 2018 2019 2020 2021

Figure 64: Incremental cost of capacity on small cell network

6.4 Outcomes

Comparing the costs of capacity on the small cell network with those typical of a macrocell
network in section 4.4.1, it is clear that the small cell network is at least an order of
magnitude cheaper – from dollar to cents or from pounds to pence. While a macrocell
would obviously not have been considered for providing coverage and capacity deep inside
a railway station, it illustrates that small cells are a highly cost-effective means of carrying
capacity. Moreover, they allow the operator to increase the density flexibly and
progressively over time in response to changing demands, without having to overbuild
capacity to an excess degree to avoid the risk of congestion.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 78
7. Our findings
In this independent report, Real Wireless has conducted an independent business case
analysis of urban small cells on behalf of Small Cell Forum.

We found that the operator market drivers for the deployment of urban small cells can be
summarised in four major areas:

 Delivering sufficient capacity, at an affordable cost, to ensure that mobile


networks keep pace with the growth in demand: while pricing and data caps
could dampen demand growth, such a strategy risks stifling the economic and
social benefits for all.
 Ensuring that coverage is available in the places where mobile users require
access. This is increasingly everywhere, including a consistent service deep within
buildings.
 Delivering a quality of experience which meets rising user expectations. Such
quality is a blend of data speeds which rise with time, consistent coverage in the
relevant locations and sufficient capacity to avoid users experiencing congestion.
 Operators could also see the business case improve if differentiated services can
also be delivered for which their customers are willing to pay more.

The driver for capacity was illustrated by a case study comparing forecasts of demand
growth with the expected growth in the quantity of spectrum available and the potential
for improved spectrum efficiency in macrocell networks. The result is a significant shortfall
in the capacity from these sources alone in dense traffic areas, so that greater spectrum
reuse via network densification – i.e. small cells – will be inevitable for most operators.

However, we found that the exact timing and quantity of small cells needed will differ
significantly amongst operators depending on their existing and planned spectrum network
assets. However, it is important to appreciate that the different sources of capacity
(spectrum, technology and topology / densification) act in combination to give greater gains
than any one source in isolation, so operators should seek advancement in all three
simultaneously.

We examined the cost effectiveness of capacity enhancement via urban small cells via a
detailed study of the costs and benefits of deploying outdoor small cells in an urban area,
with demand rising over a seven-year study period.

Our base case analysis focused on a dense urban area of 37.6 km2 with LTE traffic growing
at a compound annual growth rate of 35% over the period 2014-2020. We find that the
deployment of small cells to enhance capacity yields a benefit or $48.6m (present value) for
a total cost of ownership of $29.8 million yielding a positive net present value for operators
of $28m and a Return on Investment of 136%. This is therefore a highly cost-effective
investment with a payback period within 4 years. We consider this result conservative, as
the operator bears the costs of initially establishing the small cells sites within the study
period, but these sites will be a useful asset over the long term for the operator.

Nevertheless the specific outcomes depend on the assumed parameters and will differ
significantly amongst operators and markets. We therefore conducted a series of
sensitivity investigations to determine the robustness of the business case, outcomes of
which include:

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 79
 The value of deploying small cells increases rapidly with the growth rate of the
demand. Our base case assumed 35% annual growth but many operators face
growth at much higher levels: with a growth rate of 45% the value of small cells
increases by over 260% relative to the base case.
 In the base case, we assumed that the cost of carrying excess traffic on the
macrocells was initially $0.3/GB and increases with the density of demand to be
carried. The small cell layer remains a positive investment proposition even if the
macrocell costs are constant in the face of increasing demand (which is unlikely)
at a level as low as $0.6/GB. Other studies have suggested costs in areas that are
not heavily constrained by site availability to lie in the range $0.25-$0.7 /GB.
 Whilst placing the small cell at the centre of the hotspot maximises the demand
served, the costs of extending backhaul connectivity to this location can
sometimes outweigh the benefits. In our case study, value was maximised with
small cells around 5 meters off the ‘benefit-maximising location’. In general
though, the business case was not strongly sensitive to this factor, and was
positive over the range of offsets tried.
 The costs of small cell equipment itself are small in the total cost of ownership of
the network (approximately 20% of the overall present value of the costs) during
the study period and costs are dominated by initial site establishment costs (10-
20%) and ongoing opex (approximately 50%). A 20% price increase in the cost of
the small cell results in only a 13% reduction in the net present value of the small
cell network. This suggests that operators should seek to maximise the traffic-
carrying capabilities of the small cells deployed as the technology evolves. For
example our base case assumes that the spectrum efficiency of small cells
increases from 0.9 to 1.5 bits per second per hertz over the study period: if that
could be increased to 1.5 bits per second over the whole period then the value of
the small cell deployment increases by over 28% (to $35.8 million).

Overall we consider that the capacity enhancement based market driver is an important
one with a strong business case given carefully planned deployment and this will be the
lead driver for many operators.

We examined the role small cells can play in improving the depth of coverage provided by
an operator who does not have any low (< 1 GHz) frequency spectrum and wishes to
achieve similar coverage to their competitors. Depending on the coverage depth and
minimum data rate targeted the costs which would be incurred to match their competitor
via macrocell densification vary from $30.0m (10 metres, 35.17 Mbps) to $39.4m (15
metres, 32.24 Mbps)11. Achieving comparable coverage by deploying a dense small cell
layer across the same 37.6 km2 city area studied in the capacity enhancement study
increases the costs by between $53.8m and $55.7m. Wide area deployment of small cells is
unlikely to be as cost effective as macrocell deployment. However, if the small cells are
targeted more explicitly at the cell-edge areas where the macrocells deliver unacceptable
coverage, cost savings can be achieved compared to achieving the same target by macro
layer augmentation. These cost savings range from $2.8m to $7.2m.

There are also benefits for the operator and its customers arising from quality of
experience enhancement in the form of an increase in the median throughput delivered to
all network users (not only those on the small cells). We estimate this benefit will increase

11 Costs are presented on a present value basis.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 80
throughput by between 546% and 579% for the 68% to 37.7% of users in the area targeted
by small cells. Based on other work, we estimate this could yield a benefit ranging from
$6.3m to $11.3m for those users via improved working efficiencies and between $1.0m to
$1.9m for the operator. In addition, this coverage augmentation and performance
improvement can be achieved without requiring additional spectrum. Based on current
spectrum valuations per population served, we anticipate that reuse of the existing
spectrum in a small cell layer will save the operator just over $5m.

We examined the use of indoor urban small cells in providing capacity to indoor areas with
ultra-dense demand requirements which are often encountered in large public buildings.
We have examined this need via a case study of a busy city centre railway station. We find
that the demand density in such locations may significantly exceed the capacity density of
conventional solutions and urban small cells can play an important role in serving such
locations. We find that the cost of capacity in such a location is between 0.3-2.3 US cents
per GB, which is dramatically lower than the typical cost of macrocell capacity illustrates
that small cells are a highly cost-effective means of carrying capacity and provide flexibility
to respond progressively over time to changing usage patterns.

Overall, we concluded that urban small cells, in both outdoor and indoor incarnations are
an essential tool for operators and can yield a positive business case even when an
operator plans to evolve both macrocell capacity and spectrum holdings. However the
timing and quantity of deployment will vary markedly amongst operators who should
carefully determine the appropriate market drivers for them and take full account of
existing network and spectrum assets and demand expectations when planning their
deployments. Nevertheless, we expect that a fully-featured small cell layer will be an
essential component of future networks and all operators are advised to secure
appropriate site and backhaul assets in the near term to ensure the efficiency of future
deployments. They are also advised to consider market drivers for small cells well beyond
capacity enhancement in order to maximise their competitive position and service offering.

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 81
References
1 “Femtocell business case”, Signals Research for Small Cell Forum, SCF 005, http://scf.io
2 “The business case for enterprise small cells”, Real Wireless for Small Cell Forum, SCF062,
http://scf.io.
3 “What is the impact of mobile telephony on economic growth?” A report for the GSM
Association, Deloitte, GSMA, Cisco, November 2012
4 “Survey of operator views on the motivators and barriers for urban small cell
deployment”, Rethink Technology for Small Cell Forum, 2014, http://scf.io
5 GSMA Mobile Economy 2013, AT Kearney for GSMA.
7 “Techniques for increasing the capacity of wireless broadband networks: UK, 2012-2030″
Real Wireless for Ofcom, March 2012 http://www.realwireless.biz/mobile-capacity-in-the-
uk-major-study-published/
8 Cooper’s Law: http://en.wikipedia.org/wiki/Martin_Cooper_(inventor)#Cooper.27s_Law
9 “An assessment of the value of small cell services to operators”, Real Wireless for Virgin
Media, October 2012 http://www.realwireless.biz/small-cells-as-a-service-trials-report/
10 “W-CDMA open access small cells”, Small Cell Forum, SCF038, http://scf.io
11 “Small Cell Services”, Small Cell Forum, SCF046, http://scf.io

13 “Obama plans to double wireless spectrum”, CBC News 2010


http://www.cbc.ca/news/technology/obama-plans-to-double-wireless-spectrum-1.871247
14 UK HM Treasury Spending Review 2010,
https://www.gov.uk/government/publications/spending-review-2010
15 “RSPP: the roadmap for a wireless Europe”, European Commission 2012,
http://ec.europa.eu/digital-agenda/en/rspp-roadmap-wireless-europe
16 “Don’t worry, mobile broadband is profitable”, Ericsson, Mobile Broadband, 2009.
http://www.ericsson.com/ericsson/corpinfo/publications/ericsson_business_review/pdf/2
09/209_BUSINESS_CASE_mobile_broadband.pdf
17 “Mobile data growth – too much of a good thing?”, Plum Consulting, January 2012,
http://www.plumconsulting.co.uk/pdfs/Plum_Insight_Jan2012_Mobile_data_growth_-
_too_much_of_a_good_thing.pdf
18 “Femtocells – Natural solution for offload”, Small Cell Forum, SCF016, http://scf.io.
19 “Mobile Data Strategy”, Ofcom, November 2013,
http://stakeholders.ofcom.org.uk/binaries/consultations/mobile-data-
strategy/summary/MDS_Condoc.pdf
20 “HSDPA/HSUPA for UMTS High Speed Radio Access for Mobile Communications”, edited
by Harri Holma and Antti Toskala, John Wiley, 2006.
21 Sitefinder dataset (database of mobile sites in the UK) , accessed January 2014,
http://stakeholders.ofcom.org.uk/sitefinder/sitefinder-dataset/
22 “RW spectrum requirements for mobile broadband, a report prepared for Ofcom,
appendices to final report”, V2-0.docx Issue date: 26 June 2013.
http://stakeholders.ofcom.org.uk/binaries/consultations/cfi-mobile-
bb/annexes/RW_appendices.pdf
23 “Performance of Macro- And Co-Channel Femtocells In A Hierarchical Cell Structure”, H.
Claussen, The 18th Annual IEEE International Symposium on Personal, Indoor and Mobile
Radio Communications, 2007.
24 “LTE Advanced”, presentation by Qualcomm, February 2012, available at
http://www.qualcomm.com/media/documents/lte-advanced-global-4g-solution

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 82
25 “Deployment strategies for Heterogeneous Networks”, whitepaper by Nokia Siemens
Networks, 2012.
http://nsn.com/sites/default/files/document/nsn_deployment_strategies_for_heterogene
ous_networks_white_paper.pdf
26 “Considerations on non-CA based heterogeneous deployments”, R1-101752, Ericsson,
ST-Ericsson, 3GPP TSG RAN WG1 meeting #60bis, Beijing, China, 12 – 16 April, 2010.
http://www.3gpp.org/ftp/tsg_ran/wg1_rl1/TSGR1_60b/Docs/
27 “A Comparison of LTE Advanced HetNets and Wi-Fi”, whitepaper by Qualcomm, July
2013, http://www.qualcomm.com/media/documents/comparison-lte-advanced-hetnets-
and-wifi
28 “Heterogeneous cellular networks: From theory to practice”, A. Ghosh et. al., IEEE
Communications Magazine, June 2012
29 “Backhaul Technologies for Small Cells”, Small Cell Forum, SCF049, http://scf.io
30 “Guidelines for evaluation of radio interface technologies for IMT-Advanced” ITU-R
Report M.2135-1, 2009. http://www.itu.int/pub/R-REP-M.2135/en
31 ”Radio Frequency (RF) system scenarios”, 3GPP TR 36.942,
http://www.3gpp.org/DynaReport/36942.htm
32 “Placing Coverage and Capacity where it’s needed”, Alcatel-Lucent, http://www.alcatel-
lucent.com/wps/DocumentStreamerServlet?LMSG_CABINET=Docs_and_Resource_Ctr&LM
SG_CONTENT_FILE=Brochures/MetroCells_EN_Brochure.pdf
33 “Labour costs in the EU-27”, Eurostat, April 2013.
http://epp.eurostat.ec.europa.eu/cache/ITY_PUBLIC/3-10042013-AP/EN/3-10042013-AP-
EN.PDF
34 “Alcatel-Lucent 9364 Outdoor Metro Cell Outdoor V2”, Alcatel-Lucent data sheet,
http://www3.alcatel-
lucent.com/wps/DocumentStreamerServlet?LMSG_CABINET=Docs_and_Resource_Ctr&LM
SG_CONTENT_FILE=Data_Sheets/9364_MC_OD_V2_Eur_EN_Datasheet.pdf
35 “Prices of fuel purchased in the UK”, UK Government Department for Energy and
Climate Change, Sept 2013.
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/244619/
qep341.xls
36 “Economics of backhaul”, Exalt, http://www.exaltcom.com/Economics-of-Backhaul.aspx
37 “Leased lines charge control”, Ofcom, July 2012,
http://stakeholders.ofcom.org.uk/binaries/consultations/llcc-
2012/summary/LLCC_2012.pdf
38 “Slowdown in the rate of inflation”, UK Office of National Statistics, October 2013,
http://www.ons.gov.uk/ons/rel/cpi/consumer-price-indices/october-2013/consumer-price-
inflation-summary--october-2013.html
39 “The business case for enterprise small cells”, Real Wireless for Small Cell Forum,
SCF062, http://scf.io.
40 “HeNB (LTE Femto) network architecture”, Small Cell Forum, SCF025, http://scf.io
41 “Technical analysis of the cost of extending an 800MHz mobile broadband coverage
obligation for the United Kingdom”, Real Wireless report for Ofcom, January 2012.
http://www.realwireless.biz/800coverage/
[42] “Assessment of future mobile competition and award of 800 MHz and 2.6 GHz”,
Ofcom, July 2012,
http://stakeholders.ofcom.org.uk/binaries/consultations/award-
800mhz/statement/statement.pdf
[43] “Monte-Carlo simulation methodology for the use in sharing and compatibility studies

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 83
between different radio services or systems”, European Conference of Postal and
Telecommunications Administrations (CEPT), Report 68, June 2002.
44 Technical analysis of the cost of extending an 800 MHz mobile broadband coverage
obligation for the United Kingdom, Real Wireless report for Ofcom, January 2012.
45 “UK Broadband Impact Study”, SQW Group for Department for Culture, Media & Sport,
November 2013
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/257006/
UK_Broadband_Impact_Study_-_Impact_Report_-_Nov_2013_-_Final.pdf
46 “Daytime Population by Borough”, Greater London Authority, November 2013,
http://data.london.gov.uk/datastore/package/daytime-population-borough
47 “Public Sector Employment”, Office of National Statistics, September 2013,
http://www.ons.gov.uk/ons/rel/pse/public-sector-employment/q2-2013/sty-public-
section-employment.html
48 “Regional Gross Value Added, 2011- Interactive maps “, Office for National Statistics,
December 2012, http://www.ons.gov.uk/ons/dcp171780_291820.pdf
49 “Impact of radio spectrum on the UK economy and factors influencing future spectrum
demand”, Analysys Mason for UK Government Department of Culture, Media and Sport,
November 2012 https://www.gov.uk/government/publications/impact-of-radio-spectrum-
on-the-uk-economy-and-factors-influencing-future-spectrum-demand
50 “Spectrum value of 800MHz, 1800MHz and 2.6GHz”, DotEcon and Aetha report for
Ofcom, July 2012, http://stakeholders.ofcom.org.uk/binaries/consultations/award-
800mhz/statement/Ofcom-4G-spectrum-reserve-prices.pdf
51“Auctions Summary", GSMA, retrieved January 2014,
http://www.gsma.com/spectrum/wp-content/uploads/DigitalDividend/DDtoolkit/auctions-
summary.html
52 Think SmallCell Interview with Andy Sutton, Principal Network Architect at EE UK on
their future small cell plans http://www.thinksmallcell.com/Femtocell-
Interview/thinksmallcell-interview-with-andy-sutton-principal-network-architect-at-ee-uk-
on-their-future-small-cell-plans.html
53 “4G Capacity Gains”, Real Wireless for Ofcom, January 2011 available from
http://www.realwireless.biz/2011/05/13/ofcom-publishes-4g-capacity-gains-study-by-real-
wireless/

The Business Case for Urban Small Cells


Issue date: 04 February 2014
Version: 3.1 84
Copyright ©2014 Real Wireless Limited. All rights reserved.

Real Wireless Ltd


PO Box 2218
Pulborough t +44 207 117 8514
West Sussex f +44 808 280 0142
RH20 4XB e info@realwireless.biz
United Kingdom www.realwireless.biz

You might also like